A4

A4/TP_A4/Modele_exo4/CPP.txt

@@ -0,0 +1,42 @@
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 321.471008
+1000 features detected
+Convolution Time 0.080770
+EigenValues Time 0.321471
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 11066.154297
+1000 features detected
+Convolution Time 2.545196
+EigenValues Time 11.066154
+detecting features for ./input_images/img_2.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [2268 x 4032] px
+[gettimeofday] EigenValues Time 6031.787109
+1000 features detected
+Convolution Time 1.477949
+EigenValues Time 6.031787
+detecting features for ./input_images/img_3.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [3072 x 4608] px
+[gettimeofday] EigenValues Time 9350.551758
+1000 features detected
+Convolution Time 2.147998
+EigenValues Time 9.350552
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 10454.441406
+1000 features detected
+Convolution Time 2.435961
+EigenValues Time 10.454440
+detecting features for ./input_images/img_5.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [2656 x 3976] px
+[gettimeofday] EigenValues Time 6981.282227
+1000 features detected
+Convolution Time 1.657244
+EigenValues Time 6.981283

A4/TP_A4/Modele_exo4/modele.txt

@@ -0,0 +1,7 @@
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 50, max_features = 10000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 1587.637939
+2702 features detected
+Convolution Time 0.024775
+EigenValues Time 1.587637

A4/TP_A4/Modele_exo4/script.sh

@@ -0,0 +1,9 @@
+#!/bin/bash
+
+for s in 1 2 5 10 15 20
+do
+    for f in 5 10 50 100 500 1000
+    do
+        ./ShiTomasi ./input_images/000009.png $s $f >> time.txt
+    done
+done

A4/TP_A4/exo4_opti/CPP.txt

@@ -0,0 +1,84 @@
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 103.567001
+1000 features detected
+Convolution Time 0.122011
+EigenValues Time 0.103566
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 2251.462891
+1000 features detected
+Convolution Time 0.671026
+EigenValues Time 2.251463
+detecting features for ./input_images/img_2.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [2268 x 4032] px
+[gettimeofday] EigenValues Time 1126.833984
+1000 features detected
+Convolution Time 0.366891
+EigenValues Time 1.126834
+detecting features for ./input_images/img_3.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [3072 x 4608] px
+[gettimeofday] EigenValues Time 1776.987061
+1000 features detected
+Convolution Time 0.597440
+EigenValues Time 1.776987
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1873.168945
+1000 features detected
+Convolution Time 0.632375
+EigenValues Time 1.873169
+detecting features for ./input_images/img_5.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [2656 x 3976] px
+[gettimeofday] EigenValues Time 1474.021973
+1000 features detected
+Convolution Time 0.470415
+EigenValues Time 1.474022
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 321.471008
+1000 features detected
+Convolution Time 0.080770
+EigenValues Time 0.321471
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 11066.154297
+1000 features detected
+Convolution Time 2.545196
+EigenValues Time 11.066154
+detecting features for ./input_images/img_2.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [2268 x 4032] px
+[gettimeofday] EigenValues Time 6031.787109
+1000 features detected
+Convolution Time 1.477949
+EigenValues Time 6.031787
+detecting features for ./input_images/img_3.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [3072 x 4608] px
+[gettimeofday] EigenValues Time 9350.551758
+1000 features detected
+Convolution Time 2.147998
+EigenValues Time 9.350552
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 10454.441406
+1000 features detected
+Convolution Time 2.435961
+EigenValues Time 10.454440
+detecting features for ./input_images/img_5.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [2656 x 3976] px
+[gettimeofday] EigenValues Time 6981.282227
+1000 features detected
+Convolution Time 1.657244
+EigenValues Time 6.981283

A4/TP_A4/exo4_opti/Makefile

@@ -0,0 +1,29 @@
+CC=gcc
+CCFLAGS=-Wall -std=c99 -fopenmp
+LDFLAGS=-lm -fopenmp
+SOURCES=$(wildcard *.c)
+OBJECTS=$(SOURCES:.c=.o)
+TARGET=ShiTomasi
+
+all: debug
+
+debug: CCFLAGS += -DDEBUG -g
+debug: $(TARGET)
+
+release: CCFLAGS += -O2
+release: $(TARGET)
+
+benchmode: CCFLAGS += -O2 -DBENCHMARKMODE
+benchmode: $(TARGET)
+
+$(TARGET): $(OBJECTS) $(CXXOBJECTS)
+	$(CC) -o $@ $^ $(LDFLAGS)
+
+%.o: %.c %.h
+	$(CC) $(CCFLAGS) -c $<
+
+%.o: %.c
+	$(CC) $(CCFLAGS) -c $<
+
+clean:
+	rm -f *.pgm *.o $(TARGET)

A4/TP_A4/exo4_opti/ShiTomasi.c

@@ -0,0 +1,375 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdint.h>
+
+#include <math.h>
+#include <string.h>
+#include <sys/time.h>
+#include "image_io.h"
+#include "ShiTomasi.h"
+#include <omp.h>
+#include <sys/time.h>
+
+float timedifference_msec(struct timeval t0, struct timeval t1)
+{
+	return (t1.tv_sec - t0.tv_sec) * 1000.0f + (t1.tv_usec - t0.tv_usec) / 1000.0f;
+}
+
+int main(int argc, char **argv)
+{
+
+	// path to the image to process
+	char *filepath = NULL;
+
+	// sigma of the gaussian distribution
+	float sigma = 1.1;
+	// size of a pixel 'neighborhood'
+	int windowsize = 4;
+	// # of features
+	int max_features = 1024;
+
+	// argument parsing logic
+	if (argc == 4)
+	{
+		filepath = argv[1];
+		windowsize = atof(argv[2]);
+		max_features = atoi(argv[3]);
+	}
+	else
+	{
+		help(NULL);
+	}
+
+	printf("detecting features for %s\n", filepath);
+	printf("sigma = %0.3f, windowsize = %d, max_features = %d\n", sigma, windowsize, max_features);
+
+	int width;
+	int height;
+	int kernel_width;
+	int a;
+
+	// calculate kernel width based on sigma
+	a = (int)round(2.5 * sigma - .5);
+	kernel_width = 2 * a + 1;
+
+	// malloc and read the image to be processed
+	float *original_image;
+
+	read_imagef(filepath, &original_image, &width, &height);
+
+	printf("image_size: [%d x %d] px\n", width, height);
+
+	// malloc and generate the kernels
+	float *gkernel = (float *)malloc(sizeof(float) * kernel_width);
+	float *dkernel = (float *)malloc(sizeof(float) * kernel_width);
+	gen_kernel(gkernel, dkernel, sigma, a, kernel_width);
+
+	// create hgrad and vgrad and temp
+	float *hgrad = (float *)malloc(sizeof(float) * width * height);
+	float *vgrad = (float *)malloc(sizeof(float) * width * height);
+	float *tmp_image = (float *)malloc(sizeof(float) * width * height);
+
+	int nb_thread = omp_get_max_threads();
+	omp_set_num_threads(nb_thread);
+
+	double timer_start_conv = omp_get_wtime();
+	// convolve to get the vgrad and hgrad
+	convolve(gkernel, original_image, tmp_image, width, height, kernel_width, 1, a);
+	convolve(dkernel, tmp_image, vgrad, width, height, 1, kernel_width, a);
+	convolve(gkernel, original_image, tmp_image, width, height, 1, kernel_width, a);
+	convolve(dkernel, tmp_image, hgrad, width, height, kernel_width, 1, a);
+	
+	double timer_elapsed_conv = omp_get_wtime() - timer_start_conv;
+
+	free(tmp_image);
+	free(gkernel);
+	free(dkernel);
+
+	// Compute the eigenvalues of each pixel's z matrix. After this we can free the gradients.
+	data_wrapper_t *eigenvalues = (data_wrapper_t *)malloc(sizeof(data_wrapper_t) * width * height);
+
+	struct timeval current_time, end_time;
+	double timer_start_eigen = omp_get_wtime();
+	gettimeofday(&current_time, NULL);
+	compute_eigenvalues(hgrad, vgrad, height, width, windowsize, eigenvalues);
+	double timer_elapsed_eigen = omp_get_wtime() - timer_start_eigen;
+	gettimeofday(&end_time, NULL);
+	// float elapsed = timedifference_msec(current_time, end_time);
+	float elapsed = (end_time.tv_sec - current_time.tv_sec) * 1000.0f + (end_time.tv_usec - current_time.tv_usec) / 1000.0f;
+
+	// long double elapsed_double = (float)elapsed / (float)1000000;
+	printf("[gettimeofday] EigenValues Time %f\n", elapsed);
+	// printf("[gettimeofday] EigenValues Time %ld\n", (long)elapsed_double);
+
+	free(hgrad);
+	free(vgrad);
+
+	// Find the features based on the eigenvalues.
+	data_wrapper_t *features;
+	unsigned int features_count;
+
+	features_count = find_features(eigenvalues, max_features, width, height, &features);
+
+	free(eigenvalues);
+
+	printf("%d features detected\n", features_count);
+
+	printf("Convolution Time %f\n", timer_elapsed_conv);
+	printf("EigenValues Time %f\n", timer_elapsed_eigen);
+
+	// printf("\t");
+	// print_features(features, features_count);
+
+	// Mark the features in the output image.
+	draw_features(features, features_count, original_image, width, height);
+
+	free(features);
+
+	// Now we write the output.
+	char corner_image[30];
+	sprintf(corner_image, "output_images/corners.pgm");
+
+	write_imagef(corner_image, original_image, width, height);
+
+	// Free stuff leftover.
+	free(original_image);
+
+	return 0;
+}
+
+void draw_features(data_wrapper_t *features, unsigned int count, float *image, int image_width, int image_height)
+{
+	int radius = image_width * 0.0025;
+	//#pragma omp parallel for 
+	for (int i = 0; i < count; ++i)
+	{
+		int x = features[i].x;
+		int y = features[i].y;
+		for (int k = -1 * radius; k <= radius; k++)
+		{
+			for (int m = -1 * radius; m <= radius; m++)
+			{
+				if ((x + k) >= 0 && (x + k) < image_height && (y + m) >= 0 && (y + m) < image_width)
+					image[(x + k) * image_width + (y + m)] = 0;
+			}
+		}
+	}
+}
+
+unsigned int find_features(data_wrapper_t *eigenvalues, int max_features, int image_width, int image_height, data_wrapper_t **features)
+{
+	size_t image_size = image_height * image_width;
+
+	// Sort eigenvalues in descending order while keeping their corresponding pixel index in the image.
+	qsort(eigenvalues, image_height * image_width, sizeof *eigenvalues, sort_data_wrapper_value_desc);
+
+	// Create the features buffer based on the max_features value (acts as a percentage of the image size).
+	*features = (data_wrapper_t *)malloc(sizeof(data_wrapper_t) * max_features);
+
+	// Fill the features buffer!
+	unsigned int features_count = 0;
+	const int ignore_x = 3; // ignore this many pixels rows from top/bottom of image
+	const int ignore_y = 3; // ignore this many pixels columns from left/right of image
+	for (int i = 0; i < image_size && features_count < max_features; ++i)
+	{
+		// Ignore top left, top right, bottom right, bottom left edges of image.
+		if (eigenvalues[i].x <= ignore_x || eigenvalues[i].y <= ignore_y ||
+			eigenvalues[i].x >= image_width - 1 - ignore_x || eigenvalues[i].y >= image_height - 1 - ignore_y)
+		{
+			continue;
+		}
+
+		// Have to seed the first feature so we have a place to start.
+		if (features_count == 0)
+		{
+			(*features)[0] = eigenvalues[i];
+			features_count++;
+		}
+
+		// Check if prospective feature is more than 8 manhattan distance away from any existing feature.
+		int is_good = 1;
+		for (int j = 0; j < features_count; ++j)
+		{
+			int manhattan = abs((*features)[j].x - eigenvalues[i].x) + abs((*features)[j].y - eigenvalues[i].y);
+			if (manhattan <= 8)
+			{
+				is_good = 0;
+				break;
+			}
+		}
+
+		// If the prospective feature was at least 8 manhattan distance from all existing features, then we can add it.
+		if (is_good)
+		{
+			(*features)[features_count] = eigenvalues[i];
+			features_count++;
+		}
+	}
+
+	return features_count;
+}
+
+int sort_data_wrapper_value_desc(const void *a, const void *b)
+{
+	const data_wrapper_t *aa = (const data_wrapper_t *)a;
+	const data_wrapper_t *bb = (const data_wrapper_t *)b;
+	return (aa->data < bb->data) - (aa->data > bb->data);
+}
+
+int sort_data_wrapper_index_asc(const void *a, const void *b)
+{
+	const data_wrapper_t *aa = (const data_wrapper_t *)a;
+	const data_wrapper_t *bb = (const data_wrapper_t *)b;
+
+	if (aa->x == bb->x)
+		return ((aa->y > bb->y) - (aa->y < bb->y));
+	else
+		return ((aa->x > bb->x) - (aa->x < bb->x));
+}
+
+void compute_eigenvalues(float *hgrad, float *vgrad, int image_height, int image_width, int windowsize, data_wrapper_t *eigenvalues)
+{
+	int w = floor(windowsize / 2);
+
+	int i, j, k, m, offseti, offsetj;
+	float ixx_sum, iyy_sum, ixiy_sum;
+
+	#pragma omp parallel for collapse(2) schedule(static,100)
+	for (i = 0; i < image_height; i++)
+	{
+		for (j = 0; j < image_width; j++)
+		{
+			ixx_sum = 0;
+			iyy_sum = 0;
+			ixiy_sum = 0;
+
+			// #pragma omp parallel for reduction(+:ixx_sum) reduction(+:iyy_sum) reduction(+:ixiy_sum)
+			for (k = 0; k < windowsize; k++)
+			{
+				for (m = 0; m < windowsize; m++)
+				{
+					offseti = -1 * w + k;
+					offsetj = -1 * w + m;
+					if (i + offseti >= 0 && i + offseti < image_height && j + offsetj >= 0 && j + offsetj < image_width)
+					{
+						ixx_sum += hgrad[(i + offseti) * image_width + (j + offsetj)] * hgrad[(i + offseti) * image_width + (j + offsetj)];
+						iyy_sum += vgrad[(i + offseti) * image_width + (j + offsetj)] * vgrad[(i + offseti) * image_width + (j + offsetj)];
+						ixiy_sum += hgrad[(i + offseti) * image_width + (j + offsetj)] * vgrad[(i + offseti) * image_width + (j + offsetj)];
+					}
+				}
+			}
+
+			eigenvalues[i * image_width + j].x = i;
+			eigenvalues[i * image_width + j].y = j;
+			eigenvalues[i * image_width + j].data = min_eigenvalue(ixx_sum, ixiy_sum, ixiy_sum, iyy_sum);
+		}
+	}
+}
+
+float min_eigenvalue(float a, float b, float c, float d)
+{
+	float ev_one = (a + d) / 2 + pow(((a + d) * (a + d)) / 4 - (a * d - b * c), 0.5);
+	float ev_two = (a + d) / 2 - pow(((a + d) * (a + d)) / 4 - (a * d - b * c), 0.5);
+	if (ev_one >= ev_two)
+	{
+		return ev_two;
+	}
+	else
+	{
+		return ev_one;
+	}
+}
+
+void convolve(float *kernel, float *image, float *resultimage, int image_width, int image_height, int kernel_width, int kernel_height, int half)
+{
+	float sum;
+	int i, j, k, m, offsetj, offseti;
+	// assign the kernel to the new array
+	// #pragma omp parallel for collapse(2) 
+	#pragma omp parallel for schedule(static,100) 
+	for (i = 0; i < image_height; i++)
+	{
+		for (j = 0; j < image_width; j++)
+		{
+
+			// reset tracker
+			sum = 0.0;
+			// for each item in the kernel
+			// #pragma omp parallel for collapse(2)
+			for (k = 0; k < kernel_height; k++)
+			{
+				for (m = 0; m < kernel_width; m++)
+				{
+					offseti = -1 * (kernel_height / 2) + k;
+					offsetj = -1 * (kernel_width / 2) + m;
+					if (i + offseti >= 0 && i + offseti < image_height && j + offsetj >= 0 && j + offsetj < image_width)
+					{
+						sum += (float)(image[(i + offseti) * image_width + (j + offsetj)]) * kernel[k * kernel_width + m];
+					}
+				}
+			}
+			// copy it back
+			resultimage[i * image_width + j] = sum;
+		}
+	}
+}
+
+void gen_kernel(float *gkernel, float *dkernel, float sigma, int a, int w)
+{
+	int i;
+	float sum_gkern;
+	float sum_dkern;
+	sum_gkern = 0;
+	sum_dkern = 0;
+	// #pragma omp parallel for reduction(+:sum_gkern) reduction(-:sum_dkern) 
+	for (i = 0; i < w; i++)
+	{
+		gkernel[i] = (float)exp((float)(-1.0 * (i - a) * (i - a)) / (2 * sigma * sigma));
+		dkernel[i] = (float)(-1 * (i - a)) * (float)exp((float)(-1.0 * (i - a) * (i - a)) / (2 * sigma * sigma));
+		sum_gkern = sum_gkern + gkernel[i];
+		sum_dkern = sum_dkern - (float)i * dkernel[i];
+	}
+
+	// reverse the kernel by creating a new kernel, yes not ideal
+	float *newkernel = (float *)malloc(sizeof(float) * w);
+	// #pragma omp parallel for 
+	for (i = 0; i < w; i++)
+	{
+		dkernel[i] = dkernel[i] / sum_dkern;
+		gkernel[i] = gkernel[i] / sum_gkern;
+		newkernel[w - i] = dkernel[i];
+	}
+
+	// copy new kernel back in
+	// #pragma omp parallel for  
+	for (i = 0; i < w; i++)
+	{
+		dkernel[i] = newkernel[i + 1];
+	}
+	free(newkernel);
+}
+
+void help(const char *err)
+{
+	if (err != NULL)
+		printf("%s\n", err);
+	printf("Utilisation: ./ShiTomasi <chemin vers l'image> [Taille de la fenêtre] [Nombre de primitives] \n");
+	printf("arguments:\n");
+	printf("\tTaille de la fenêtre: taille du voisinage d'un pixel dans l'image\n");
+	printf("\tNombre de primitives: nombre de primitives à extraire\n");
+	exit(0);
+}
+
+void print_features(data_wrapper_t *features, unsigned int count)
+{
+	// Sort the features by
+	qsort(features, count, sizeof *features, sort_data_wrapper_index_asc);
+	for (unsigned int i = 0; i < count; ++i)
+	{
+		if (i % 15 != 0 || i == 0)
+			printf("(%d,%d) ", features[i].x, features[i].y);
+		else
+			printf("(%d,%d)\n\t", features[i].x, features[i].y);
+	}
+	printf("\n");
+}

A4/TP_A4/exo4_opti/ShiTomasi.h

@@ -0,0 +1,41 @@
+#ifndef SHI_TOMASI_H
+#define SHI_TOMASI_H
+
+typedef struct data_wrapper_t
+{
+	float data;
+	int x;
+	int y;
+} data_wrapper_t;
+
+/// Draws a box at specfied location in the image. Used for markgin features.
+void draw_features(data_wrapper_t *features, unsigned int count, float *image, int image_width, int image_height);
+
+/// Find features in an image.
+unsigned int find_features(data_wrapper_t *eigenvalues, int max_features, int image_width, int image_height, data_wrapper_t **features);
+
+/// Defines comparison for data_wrapper_t. When used with qsort, it will result in a descending order array.
+int sort_data_wrapper_value_desc(const void *a, const void *b);
+
+// Sort data_wrapper_t types by their index (x first then y) in ascending order.
+int sort_data_wrapper_index_asc(const void *a, const void *b);
+
+/// Compute the eigenvalues of a pixel's Z matrix.
+void compute_eigenvalues(float *hgrad, float *vgrad, int image_height, int image_width, int windowsize, data_wrapper_t *eigenvalues);
+
+/// Calculate the minimum eigenvalue.
+float min_eigenvalue(float a, float b, float c, float d);
+
+/// Produce the images horizontal and vertical gradients.
+void convolve(float *kernel, float *image, float *resultimage, int image_width, int image_height, int kernel_width, int kernel_height, int half);
+
+/// Creates Gaussian kernel and Gaussian derivative kernel for image gradient/convolution procedure.
+void gen_kernel(float *gkernel, float *dkernel, float sigma, int a, int w);
+
+/// Prints out the program help menu.
+void help(const char *err);
+
+/// Print out all of the detected features.
+void print_features(data_wrapper_t *features, unsigned int count);
+
+#endif

A4/TP_A4/exo4_opti/graph.py

@@ -0,0 +1,75 @@
+import numpy as np
+import re
+from matplotlib import pyplot as plt
+
+### Variation de W et E
+
+# with open ('time.txt', 'r') as timing:
+#     text = timing.read()
+
+# # extract the values using regular expression
+# values = re.findall(r'\d+\.\d+|\d+', text)
+
+# # convert values to int or float
+# values = [float(value) if '.' in value else int(value) for value in values]
+
+# result = np.reshape(values,(36,10))
+# # print(result)
+# # SEQ = result[0:6,3]
+
+# result = np.delete(result, (0,1,4,5,6,7), 1)
+
+
+# W1 = result[0:6,:]
+# W2 = result[6:12,:]
+# W5 = result[12:18,:]
+# W10 = result[18:24,:]
+# W15 = result[24:30,:]
+# W20 = result[30:36,:]
+
+# plt.figure()
+# plt.plot(W1[:,1],W1[:,2]+W1[:,3],'+-')
+# plt.plot(W2[:,1],W2[:,2]+W2[:,3],'+-')
+# plt.plot(W5[:,1],W5[:,2]+W5[:,3],'+-')
+# plt.plot(W10[:,1],W10[:,2]+W10[:,3],'+-')
+# plt.plot(W15[:,1],W15[:,2]+W15[:,3],'+-')
+# plt.plot(W20[:,1],W20[:,2]+W20[:,3],'+-')
+
+# plt.legend(["WS = 1","WS = 2","WS = 5","WS = 10","WS = 15","WS = 20"])
+# plt.xlabel("Coins")
+# plt.ylabel("t (s)")
+# plt.title("Convolution + Eigen pas opti")
+# plt.show()
+
+
+### CPP
+
+with open ('CPP.txt', 'r') as timing:
+    text = timing.read()
+
+# extract the values using regular expression
+values = re.findall(r'\d+\.\d+|\d+', text)
+
+# convert values to int or float
+values = [float(value) if '.' in value else int(value) for value in values]
+
+result = np.reshape(values,(12,10))
+# print(result)
+
+RES = result[:6,4:6]
+X = [str(sub) for sub in RES]
+
+result = np.delete(result, (0,1,4,5,6,7), 1)
+
+OPT = result[0:6,2]+result[0:6,3]*1.1e9/(RES[:,0]*RES[:,1])
+NOPT = result[6:12,2]+result[6:12,3]*1.1e9/(RES[:,0]*RES[:,1])
+
+plt.figure()
+plt.plot(X,OPT,'+-')
+plt.plot(X,NOPT,'+-')
+
+plt.legend(["Optimisé","Non optimisé"])
+plt.xlabel("Résolution")
+plt.ylabel("CPP")
+plt.title("CPP en fonction des images")
+plt.show()

A4/TP_A4/exo4_opti/image_io.c

@@ -0,0 +1,77 @@
+
+#include "image_io.h"
+#define STB_IMAGE_IMPLEMENTATION
+#include "stb_image.h"
+
+#define BUFFER 512
+
+#define READ_IMAGE_TEMPLATE(T)                                                       \
+	{                                                                                \
+		int im_channels;                                                             \
+		unsigned char *data = stbi_load(name, im_width, im_height, &im_channels, 1); \
+		if (data == NULL)                                                            \
+		{                                                                            \
+			printf("ERROR: Cannot read %s\n", name);                                 \
+			exit(0);                                                                 \
+		}                                                                            \
+		*image = malloc(sizeof(**image) * (*im_width) * (*im_height));               \
+		for (int i = 0; i < (*im_width) * (*im_height); ++i)                         \
+			(*image)[i] = (T)data[i];                                                \
+		stbi_image_free(data);                                                       \
+	}
+
+#define WRITE_IMAGE_TEMPLATE(T)                                                         \
+	{                                                                                   \
+		unsigned char *temp_img = malloc(sizeof(unsigned char) * im_width * im_height); \
+		for (int i = 0; i < (im_width * im_height); i++)                                \
+			temp_img[i] = image[i];                                                     \
+		write_imagec(name, temp_img, im_width, im_height);                              \
+		free(temp_img);                                                                 \
+	}
+
+void read_image(char *name, double **image, int *im_width, int *im_height)
+{
+	READ_IMAGE_TEMPLATE(double)
+}
+
+void read_imagef(char *name, float **image, int *im_width, int *im_height)
+{
+	READ_IMAGE_TEMPLATE(float)
+}
+
+void read_imagei(char *name, int **image, int *im_width, int *im_height)
+{
+	READ_IMAGE_TEMPLATE(int)
+}
+
+void read_imagec(char *name, unsigned char **image, int *im_width, int *im_height)
+{
+	READ_IMAGE_TEMPLATE(char)
+}
+
+void write_image(char *name, double *image, int im_width, int im_height)
+{
+	WRITE_IMAGE_TEMPLATE(double)
+}
+
+void write_imagef(char *name, float *image, int im_width, int im_height)
+{
+	WRITE_IMAGE_TEMPLATE(float)
+}
+
+void write_imagei(char *name, int *image, int im_width, int im_height)
+{
+	WRITE_IMAGE_TEMPLATE(int)
+}
+
+void write_imagec(char *name, unsigned char *image, int im_width, int im_height)
+{
+	FILE *fop;
+	int im_size = im_width * im_height;
+
+	fop = fopen(name, "w+");
+	fprintf(fop, "P5\n%d %d\n255\n", im_width, im_height);
+	fwrite(image, sizeof(unsigned char), im_size, fop);
+
+	fclose(fop);
+}

A4/TP_A4/exo4_opti/image_io.h

@@ -0,0 +1,11 @@
+#ifndef IMAGE_IO_H
+#define IMAGE_IO_H
+
+void read_image(char *name, double **image, int *im_width, int *im_height);
+void read_imagef(char *name, float **image, int *im_width, int *im_height);
+void read_imagec(char *name, unsigned char **image, int *im_width, int *im_height);
+void write_image(char *name, double *image, int im_width, int im_height);
+void write_imagef(char *name, float *image, int im_width, int im_height);
+void write_imagec(char *name, unsigned char *image, int im_width, int im_height);
+
+#endif

A4/TP_A4/exo4_opti/opti.txt

@@ -0,0 +1,48 @@
+Eigen #pragma omp parallel for collapse(2) schedule(static,100)
+-------------------------
+Convolve // #pragma omp parallel for schedule(static,100)
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 50, max_features = 10000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 306.903992
+2702 features detected
+Convolution Time 0.012532
+EigenValues Time 0.306904
+
+Convolve sans rien
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 50, max_features = 10000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 327.713989
+2702 features detected
+Convolution Time 0.018599
+EigenValues Time 0.327714
+
+Convolve // #pragma omp parallel for schedule(static,50)
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 50, max_features = 10000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 296.431000
+2702 features detected
+Convolution Time 0.037112
+EigenValues Time 0.296431
+
+-------------------------
+Convolve // #pragma omp parallel for schedule(static,200)
+-------------------------
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 50, max_features = 10000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 307.218994
+2702 features detected
+Convolution Time 0.024083
+EigenValues Time 0.307218
+
+
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 50, max_features = 10000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 306.039001
+2702 features detected
+Convolution Time 0.057235
+EigenValues Time 0.306040

A4/TP_A4/exo4_opti/script.sh

@@ -0,0 +1,9 @@
+#!/bin/bash
+
+for s in 1 2 5 10 15 20
+do
+    for f in 5 10 50 100 500 1000
+    do
+        ./ShiTomasi ./input_images/img_4.jpg $s $f >> time_opti_img4.txt
+    done
+done

A4/TP_A4/exo4_opti/time.txt

@@ -0,0 +1,252 @@
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 1, max_features = 5
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 78.111000
+5 features detected
+Convolution Time 0.071285
+EigenValues Time 0.078110
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 1, max_features = 10
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 80.944000
+10 features detected
+Convolution Time 0.073226
+EigenValues Time 0.080944
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 1, max_features = 50
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 72.806000
+50 features detected
+Convolution Time 0.071076
+EigenValues Time 0.072805
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 1, max_features = 100
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 76.814003
+100 features detected
+Convolution Time 0.074288
+EigenValues Time 0.076813
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 1, max_features = 500
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 74.902000
+500 features detected
+Convolution Time 0.073242
+EigenValues Time 0.074902
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 1, max_features = 1000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 72.913002
+1000 features detected
+Convolution Time 0.077598
+EigenValues Time 0.072913
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 2, max_features = 5
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 80.535004
+5 features detected
+Convolution Time 0.071336
+EigenValues Time 0.080535
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 2, max_features = 10
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 82.468002
+10 features detected
+Convolution Time 0.071284
+EigenValues Time 0.082467
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 2, max_features = 50
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 87.601997
+50 features detected
+Convolution Time 0.079019
+EigenValues Time 0.087602
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 2, max_features = 100
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 84.942001
+100 features detected
+Convolution Time 0.078346
+EigenValues Time 0.084942
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 2, max_features = 500
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 80.504997
+500 features detected
+Convolution Time 0.071112
+EigenValues Time 0.080505
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 2, max_features = 1000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 80.972000
+1000 features detected
+Convolution Time 0.071165
+EigenValues Time 0.080972
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 5, max_features = 5
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 130.117004
+5 features detected
+Convolution Time 0.071212
+EigenValues Time 0.130117
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 5, max_features = 10
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 130.608994
+10 features detected
+Convolution Time 0.071592
+EigenValues Time 0.130609
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 5, max_features = 50
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 133.570999
+50 features detected
+Convolution Time 0.072458
+EigenValues Time 0.133570
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 5, max_features = 100
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 130.069000
+100 features detected
+Convolution Time 0.071585
+EigenValues Time 0.130069
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 5, max_features = 500
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 131.382996
+500 features detected
+Convolution Time 0.071034
+EigenValues Time 0.131383
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 5, max_features = 1000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 130.604996
+1000 features detected
+Convolution Time 0.071517
+EigenValues Time 0.130605
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 10, max_features = 5
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 310.023987
+5 features detected
+Convolution Time 0.072872
+EigenValues Time 0.310024
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 10, max_features = 10
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 312.074005
+10 features detected
+Convolution Time 0.073316
+EigenValues Time 0.312074
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 10, max_features = 50
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 323.491028
+50 features detected
+Convolution Time 0.073598
+EigenValues Time 0.323491
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 10, max_features = 100
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 312.666992
+100 features detected
+Convolution Time 0.079026
+EigenValues Time 0.312667
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 10, max_features = 500
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 306.220001
+500 features detected
+Convolution Time 0.071294
+EigenValues Time 0.306220
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 306.755981
+1000 features detected
+Convolution Time 0.071774
+EigenValues Time 0.306757
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 15, max_features = 5
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 585.828003
+5 features detected
+Convolution Time 0.071077
+EigenValues Time 0.585827
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 15, max_features = 10
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 586.400024
+10 features detected
+Convolution Time 0.071432
+EigenValues Time 0.586399
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 15, max_features = 50
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 585.309998
+50 features detected
+Convolution Time 0.071168
+EigenValues Time 0.585310
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 15, max_features = 100
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 584.059021
+100 features detected
+Convolution Time 0.071715
+EigenValues Time 0.584059
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 15, max_features = 500
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 584.741028
+500 features detected
+Convolution Time 0.071440
+EigenValues Time 0.584741
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 15, max_features = 1000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 581.815979
+1000 features detected
+Convolution Time 0.071520
+EigenValues Time 0.581816
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 20, max_features = 5
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 1059.534058
+5 features detected
+Convolution Time 0.072195
+EigenValues Time 1.059534
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 20, max_features = 10
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 963.630005
+10 features detected
+Convolution Time 0.071242
+EigenValues Time 0.963630
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 20, max_features = 50
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 982.309998
+50 features detected
+Convolution Time 0.071566
+EigenValues Time 0.982309
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 20, max_features = 100
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 1015.593994
+100 features detected
+Convolution Time 0.070913
+EigenValues Time 1.015593
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 20, max_features = 500
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 987.187988
+500 features detected
+Convolution Time 0.070887
+EigenValues Time 0.987187
+detecting features for ./input_images/000009.png
+sigma = 1.100, windowsize = 20, max_features = 1000
+image_size: [1241 x 376] px
+[gettimeofday] EigenValues Time 988.111023
+1000 features detected
+Convolution Time 0.072037
+EigenValues Time 0.988112

A4/TP_A4/exo4_opti/time_opti_img1.txt

@@ -0,0 +1,252 @@
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 1, max_features = 5
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 2035.822021
+5 features detected
+Convolution Time 0.743961
+EigenValues Time 2.035822
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 1, max_features = 10
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 1681.629028
+10 features detected
+Convolution Time 0.707745
+EigenValues Time 1.681629
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 1, max_features = 50
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 1685.898926
+50 features detected
+Convolution Time 0.694751
+EigenValues Time 1.685899
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 1, max_features = 100
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 2294.447998
+100 features detected
+Convolution Time 0.671624
+EigenValues Time 2.294449
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 1, max_features = 500
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 1678.113037
+500 features detected
+Convolution Time 0.677528
+EigenValues Time 1.678113
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 1, max_features = 1000
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 2151.630859
+1000 features detected
+Convolution Time 0.676732
+EigenValues Time 2.151631
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 2, max_features = 5
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 1132.209961
+5 features detected
+Convolution Time 0.718599
+EigenValues Time 1.132210
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 2, max_features = 10
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 2359.069092
+10 features detected
+Convolution Time 0.787626
+EigenValues Time 2.359069
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 2, max_features = 50
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 1647.644043
+50 features detected
+Convolution Time 0.684985
+EigenValues Time 1.647644
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 2, max_features = 100
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 1073.688965
+100 features detected
+Convolution Time 0.696173
+EigenValues Time 1.073688
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 2, max_features = 500
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 1103.389038
+500 features detected
+Convolution Time 0.701647
+EigenValues Time 1.103390
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 2, max_features = 1000
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 1183.239990
+1000 features detected
+Convolution Time 0.683521
+EigenValues Time 1.183240
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 5, max_features = 5
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 1262.051025
+5 features detected
+Convolution Time 0.690352
+EigenValues Time 1.262051
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 5, max_features = 10
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 2421.749023
+10 features detected
+Convolution Time 0.732796
+EigenValues Time 2.421749
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 5, max_features = 50
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 1284.166016
+50 features detected
+Convolution Time 0.695343
+EigenValues Time 1.284166
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 5, max_features = 100
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 2460.414062
+100 features detected
+Convolution Time 0.713780
+EigenValues Time 2.460414
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 5, max_features = 500
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 1168.463989
+500 features detected
+Convolution Time 0.744367
+EigenValues Time 1.168464
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 5, max_features = 1000
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 1681.278076
+1000 features detected
+Convolution Time 0.694903
+EigenValues Time 1.681278
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 10, max_features = 5
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 2089.882080
+5 features detected
+Convolution Time 0.709452
+EigenValues Time 2.089882
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 10, max_features = 10
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 2231.855957
+10 features detected
+Convolution Time 0.688733
+EigenValues Time 2.231856
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 10, max_features = 50
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 2240.367920
+50 features detected
+Convolution Time 0.743735
+EigenValues Time 2.240368
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 10, max_features = 100
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 2052.343018
+100 features detected
+Convolution Time 0.718175
+EigenValues Time 2.052343
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 10, max_features = 500
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 2114.470947
+500 features detected
+Convolution Time 0.711941
+EigenValues Time 2.114471
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 2238.066895
+1000 features detected
+Convolution Time 0.715894
+EigenValues Time 2.238067
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 15, max_features = 5
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 4002.260986
+5 features detected
+Convolution Time 0.709723
+EigenValues Time 4.002261
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 15, max_features = 10
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 4062.758057
+10 features detected
+Convolution Time 0.684714
+EigenValues Time 4.062757
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 15, max_features = 50
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 4009.768066
+50 features detected
+Convolution Time 0.707542
+EigenValues Time 4.009768
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 15, max_features = 100
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 3917.907959
+100 features detected
+Convolution Time 0.729426
+EigenValues Time 3.917908
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 15, max_features = 500
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 3936.593018
+500 features detected
+Convolution Time 0.709724
+EigenValues Time 3.936593
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 15, max_features = 1000
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 4052.041992
+1000 features detected
+Convolution Time 0.673883
+EigenValues Time 4.052041
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 20, max_features = 5
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 6904.195801
+5 features detected
+Convolution Time 0.679914
+EigenValues Time 6.904195
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 20, max_features = 10
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 6609.250000
+10 features detected
+Convolution Time 0.707525
+EigenValues Time 6.609250
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 20, max_features = 50
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 6599.893066
+50 features detected
+Convolution Time 0.706646
+EigenValues Time 6.599893
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 20, max_features = 100
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 6917.371094
+100 features detected
+Convolution Time 0.707266
+EigenValues Time 6.917372
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 20, max_features = 500
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 6863.350098
+500 features detected
+Convolution Time 0.715015
+EigenValues Time 6.863350
+detecting features for ./input_images/img_1.jpg
+sigma = 1.100, windowsize = 20, max_features = 1000
+image_size: [3282 x 5114] px
+[gettimeofday] EigenValues Time 6598.800781
+1000 features detected
+Convolution Time 0.685112
+EigenValues Time 6.598801

A4/TP_A4/exo4_opti/time_opti_img4.txt

@@ -0,0 +1,252 @@
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 1, max_features = 5
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1155.407959
+5 features detected
+Convolution Time 0.735920
+EigenValues Time 1.155408
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 1, max_features = 10
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 2209.542969
+10 features detected
+Convolution Time 0.617611
+EigenValues Time 2.209543
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 1, max_features = 50
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 2115.429932
+50 features detected
+Convolution Time 0.661218
+EigenValues Time 2.115431
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 1, max_features = 100
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 2141.304932
+100 features detected
+Convolution Time 0.625168
+EigenValues Time 2.141304
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 1, max_features = 500
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1641.547974
+500 features detected
+Convolution Time 0.623114
+EigenValues Time 1.641547
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 1, max_features = 1000
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1058.590942
+1000 features detected
+Convolution Time 0.626698
+EigenValues Time 1.058591
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 2, max_features = 5
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1120.083984
+5 features detected
+Convolution Time 0.631218
+EigenValues Time 1.120084
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 2, max_features = 10
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1578.343018
+10 features detected
+Convolution Time 0.634375
+EigenValues Time 1.578344
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 2, max_features = 50
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1587.696045
+50 features detected
+Convolution Time 0.661663
+EigenValues Time 1.587696
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 2, max_features = 100
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1089.848022
+100 features detected
+Convolution Time 0.632955
+EigenValues Time 1.089848
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 2, max_features = 500
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1117.678955
+500 features detected
+Convolution Time 0.629164
+EigenValues Time 1.117679
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 2, max_features = 1000
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1625.005005
+1000 features detected
+Convolution Time 0.661020
+EigenValues Time 1.625006
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 5, max_features = 5
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1147.598022
+5 features detected
+Convolution Time 0.626361
+EigenValues Time 1.147599
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 5, max_features = 10
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 2099.885010
+10 features detected
+Convolution Time 0.632201
+EigenValues Time 2.099884
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 5, max_features = 50
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 2346.951904
+50 features detected
+Convolution Time 0.615773
+EigenValues Time 2.346952
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 5, max_features = 100
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 2178.379883
+100 features detected
+Convolution Time 0.640185
+EigenValues Time 2.178380
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 5, max_features = 500
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 2249.246094
+500 features detected
+Convolution Time 0.620186
+EigenValues Time 2.249246
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 5, max_features = 1000
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1252.310059
+1000 features detected
+Convolution Time 0.623934
+EigenValues Time 1.252310
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 10, max_features = 5
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1980.151978
+5 features detected
+Convolution Time 0.648045
+EigenValues Time 1.980152
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 10, max_features = 10
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1949.417969
+10 features detected
+Convolution Time 0.626475
+EigenValues Time 1.949418
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 10, max_features = 50
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 2000.031982
+50 features detected
+Convolution Time 0.620183
+EigenValues Time 2.000032
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 10, max_features = 100
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1910.555054
+100 features detected
+Convolution Time 0.651491
+EigenValues Time 1.910555
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 10, max_features = 500
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1991.965942
+500 features detected
+Convolution Time 0.640609
+EigenValues Time 1.991965
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 10, max_features = 1000
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 1924.969971
+1000 features detected
+Convolution Time 0.645218
+EigenValues Time 1.924971
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 15, max_features = 5
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 3823.891113
+5 features detected
+Convolution Time 0.639121
+EigenValues Time 3.823891
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 15, max_features = 10
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 3777.537109
+10 features detected
+Convolution Time 0.628942
+EigenValues Time 3.777537
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 15, max_features = 50
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 3737.520996
+50 features detected
+Convolution Time 0.630005
+EigenValues Time 3.737521
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 15, max_features = 100
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 3869.402100
+100 features detected
+Convolution Time 0.628104
+EigenValues Time 3.869401
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 15, max_features = 500
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 3872.395996
+500 features detected
+Convolution Time 0.653086
+EigenValues Time 3.872396
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 15, max_features = 1000
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 3690.702148
+1000 features detected
+Convolution Time 0.627411
+EigenValues Time 3.690702
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 20, max_features = 5
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 6369.575195
+5 features detected
+Convolution Time 0.616158
+EigenValues Time 6.369575
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 20, max_features = 10
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 6290.769043
+10 features detected
+Convolution Time 0.651428
+EigenValues Time 6.290769
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 20, max_features = 50
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 6509.250000
+50 features detected
+Convolution Time 0.633383
+EigenValues Time 6.509250
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 20, max_features = 100
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 6335.139160
+100 features detected
+Convolution Time 0.660637
+EigenValues Time 6.335139
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 20, max_features = 500
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 6370.258789
+500 features detected
+Convolution Time 0.622169
+EigenValues Time 6.370259
+detecting features for ./input_images/img_4.jpg
+sigma = 1.100, windowsize = 20, max_features = 1000
+image_size: [3999 x 3999] px
+[gettimeofday] EigenValues Time 6391.324219
+1000 features detected
+Convolution Time 0.645403
+EigenValues Time 6.391324