SFND 2D Feature Tracking

The idea of the camera course is to build a collision detection system - that's the overall goal for the Final Project. As a preparation for this, you will now build the feature tracking part and test various detector / descriptor combinations to see which ones perform best. This mid-term project consists of four parts:

First, you will focus on loading images, setting up data structures and putting everything into a ring buffer to optimize memory load.
Then, you will integrate several keypoint detectors such as HARRIS, FAST, BRISK and SIFT and compare them with regard to number of keypoints and speed.
In the next part, you will then focus on descriptor extraction and matching using brute force and also the FLANN approach we discussed in the previous lesson.
In the last part, once the code framework is complete, you will test the various algorithms in different combinations and compare them with regard to some performance measures.

See the classroom instruction and code comments for more details on each of these parts. Once you are finished with this project, the keypoint matching part will be set up and you can proceed to the next lesson, where the focus is on integrating Lidar points and on object detection using deep-learning.

Dependencies for Running Locally

cmake >= 2.8
- All OSes: click here for installation instructions
make >= 4.1 (Linux, Mac), 3.81 (Windows)
- Linux: make is installed by default on most Linux distros
- Mac: install Xcode command line tools to get make
- Windows: Click here for installation instructions
OpenCV >= 4.1
- This must be compiled from source using the -D OPENCV_ENABLE_NONFREE=ON cmake flag for testing the SIFT and SURF detectors.
- The OpenCV 4.1.0 source code can be found here
gcc/g++ >= 5.4
- Linux: gcc / g++ is installed by default on most Linux distros
- Mac: same deal as make - install Xcode command line tools
- Windows: recommend using MinGW

Basic Build Instructions

Clone this repo.
Make a build directory in the top level directory: mkdir build && cd build
Compile: cmake .. && make
Run it: ./2D_feature_tracking.

Code

I added code for variuos opencv algorithms for - keypoint detectors, descriptor, and matchers from two consecutive images.
I used KNN match selection (k=2) and performed descriptor distance ratio filtering with t=0.8 in file matching2D.cpp.
Created test results by running combination of algorithms for - keypoint detectors, descriptor, and matchers against KITTI images.

Benchmark

audit log for m looping multiple options and collect experiment data

Added new code to capture KPI, metrics from various experiments (with combinations of alogrithms detectors, descriptors and match algorithms).

Used 'struct' to keep track of experiments - Config2DFeatTrack and AuditLog

    
    //struct to hold experiment configuration
    struct Config2DFeatTrack {
        std::string detectorType = "SHITOMASI";
        std::string descriptorType = "BRISK"; // BRIEF, ORB, FREAK, AKAZE, SIFT
    
        std::string matcherType = "MAT_BF";        // MAT_BF, MAT_FLANN
        std::string matcherTypeMetric = "DES_BINARY"; // DES_BINARY, DES_HOG
        std::string matcherTypeSelector = "SEL_NN";       // SEL_NN, SEL_KNN
    
        bool bVis = false;
        bool bLimitKpts = false;
        int maxKeypoints = 50;
    }; 

    //struct to hold experiment audit log
    struct AuditLog {
        Config2DFeatTrack config ;
        std::string image_name ="";
        bool isError = false;
        long match_time = 0;
        long match_keypoints_size = 0;
        long match_removed_keypoints_size = 0;
        long desc_time  = 0;
        long detect_time = 0;
        long detect_keypoints_size = 0;
    };

Now dyanamically create a new array of Config2DFeatTrack with all possible combinations from following setting.

    vector<Config2DFeatTrack> configList;

    vector<string> detectorTypes = {"SHITOMASI", "HARRIS", "FAST", "BRISK", "ORB", "AKAZE", "SIFT"};
    vector<string> descriptorTypes = {"BRISK", "BRIEF", "ORB", "FREAK", "AKAZE", "SIFT"}; 
    vector<string> matcherTypes = {"MAT_BF", "MAT_FLANN"};
    vector<string> matcherTypeMetrics = {"DES_BINARY", "DES_HOG"};
    vector<string> matcherTypeSelectors = {"SEL_NN", "SEL_KNN"};

    ...
    ...
    for (auto matcherType:matcherTypes) {
        for (auto matcherTypeMetric:matcherTypeMetrics) {
            for (auto matcherTypeSelector:matcherTypeSelectors) {
                Config2DFeatTrack config;

                config.detectorType = detectorType;
                config.descriptorType = descriptorType;
                config.matcherType = matcherType;
                config.matcherTypeMetric = matcherTypeMetric;
                config.matcherTypeSelector = matcherTypeSelector;

                configList.push_back(config);
            }
        }
    }
    ...
    ...

To run all combinations of algorithms or one set of algorithm. Please change 'singleTest' in 'MidTermProject_Camera_Student.cpp -> main'

   //'MidTermProject_Camera_Student.cpp -> main'
    // for testing one set of algorithms use singleTest = true
   //if singleTest = false,  will run all combinations of algorithms
   bool singleTest = false;

   vector<Config2DFeatTrack> configList = getConfig(singleTest);

   ```

output of log outputed into three files runtest.log

results.csv

    error,image_name,detectorType,descriptorType,matcherType,matcherTypeMetric,matcherTypeSelector,detect_time,desc_time,match_time,detect_keypoints_size,match_keypoints_size,match_removed_keypoints_size,bVis,bLimitKpts,maxKeypoints
    0,../images/KITTI/2011_09_26/image_00/data/0000000000.png,SHITOMASI,BRISK,MAT_BF,DES_BINARY,SEL_NN,19,345,0,1370,0,0,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000001.png,SHITOMASI,BRISK,MAT_BF,DES_BINARY,SEL_NN,19,346,0,1301,125,0,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000002.png,SHITOMASI,BRISK,MAT_BF,DES_BINARY,SEL_NN,19,339,0,1361,118,0,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000003.png,SHITOMASI,BRISK,MAT_BF,DES_BINARY,SEL_NN,16,338,0,1358,123,0,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000004.png,SHITOMASI,BRISK,MAT_BF,DES_BINARY,SEL_NN,16,336,0,1333,120,0,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000005.png,SHITOMASI,BRISK,MAT_BF,DES_BINARY,SEL_NN,17,338,0,1284,120,0,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000006.png,SHITOMASI,BRISK,MAT_BF,DES_BINARY,SEL_NN,17,336,0,1322,113,0,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000007.png,SHITOMASI,BRISK,MAT_BF,DES_BINARY,SEL_NN,16,337,0,1366,114,0,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000008.png,SHITOMASI,BRISK,MAT_BF,DES_BINARY,SEL_NN,17,336,0,1389,123,0,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000009.png,SHITOMASI,BRISK,MAT_BF,DES_BINARY,SEL_NN,17,337,0,1339,111,0,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000000.png,SHITOMASI,BRISK,MAT_BF,DES_BINARY,SEL_KNN,16,336,0,1370,0,0,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000001.png,SHITOMASI,BRISK,MAT_BF,DES_BINARY,SEL_KNN,17,334,0,1301,95,30,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000002.png,SHITOMASI,BRISK,MAT_BF,DES_BINARY,SEL_KNN,17,334,0,1361,88,30,0,0,50
     
    0,../images/KITTI/2011_09_26/image_00/data/0000000007.png,SHITOMASI,SIFT,MAT_FLANN,DES_BINARY,SEL_KNN,11,17,3,1366,96,18,0,0,50  
    0,../images/KITTI/2011_09_26/image_00/data/0000000008.png,SHITOMASI,SIFT,MAT_FLANN,DES_BINARY,SEL_KNN,11,19,2,1389,106,17,0,0,50  
    0,../images/KITTI/2011_09_26/image_00/data/0000000009.png,SHITOMASI,SIFT,MAT_FLANN,DES_BINARY,SEL_KNN,17,15,2,1339,97,14,0,0,50  
    0,../images/KITTI/2011_09_26/image_00/data/0000000000.png,SHITOMASI,SIFT,MAT_FLANN,DES_HOG,SEL_NN,11,19,0,1370,0,0,0,0,50  
    0,../images/KITTI/2011_09_26/image_00/data/0000000001.png,SHITOMASI,SIFT,MAT_FLANN,DES_HOG,SEL_NN,11,16,2,1301,125,0,0,0,50  
    0,../images/KITTI/2011_09_26/image_00/data/0000000002.png,SHITOMASI,SIFT,MAT_FLANN,DES_HOG,SEL_NN,11,15,2,1361,118,0,0,0,50  
    0,../images/KITTI/2011_09_26/image_00/data/0000000003.png,SHITOMASI,SIFT,MAT_FLANN,DES_HOG,SEL_NN,16,15,2,1358,123,0,0,0,50  
    0,../images/KITTI/2011_09_26/image_00/data/0000000004.png,SHITOMASI,SIFT,MAT_FLANN,DES_HOG,SEL_NN,11,15,2,1333,120,0,0,0,50  
    0,../images/KITTI/2011_09_26/image_00/data/0000000005.png,SHITOMASI,SIFT,MAT_FLANN,DES_HOG,SEL_NN,12,16,2,1284,120,0,0,0,50  
    0,../images/KITTI/2011_09_26/image_00/data/0000000006.png,SHITOMASI,SIFT,MAT_FLANN,DES_HOG,SEL_NN,11,15,1,1322,113,0,0,0,50  

    0,../images/KITTI/2011_09_26/image_00/data/0000000000.png,FAST,BRISK,MAT_BF,DES_BINARY,SEL_KNN,1,327,0,1824,0,0,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000001.png,FAST,BRISK,MAT_BF,DES_BINARY,SEL_KNN,1,328,0,1832,97,52,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000002.png,FAST,BRISK,MAT_BF,DES_BINARY,SEL_KNN,1,329,0,1810,104,48,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000003.png,FAST,BRISK,MAT_BF,DES_BINARY,SEL_KNN,1,334,0,1817,101,49,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000004.png,FAST,BRISK,MAT_BF,DES_BINARY,SEL_KNN,1,328,0,1793,98,57,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000005.png,FAST,BRISK,MAT_BF,DES_BINARY,SEL_KNN,1,332,0,1796,85,64,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000006.png,FAST,BRISK,MAT_BF,DES_BINARY,SEL_KNN,1,329,0,1788,107,42,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000007.png,FAST,BRISK,MAT_BF,DES_BINARY,SEL_KNN,1,327,0,1695,107,49,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000008.png,FAST,BRISK,MAT_BF,DES_BINARY,SEL_KNN,1,325,0,1749,100,50,0,0,50
    0,../images/KITTI/2011_09_26/image_00/data/0000000009.png,FAST,BRISK,MAT_BF,DES_BINARY,SEL_KNN,1,326,0,1770,100,38,0,0,50

results.json

       [
         {
           'isError': '0',
           'image_name': '../images/KITTI/2011_09_26/image_00/data/0000000000.png',
           'detectorType': 'SHITOMASI',
           'descriptorType': 'BRISK',
           'matcherType': 'MAT_BF',
           'matcherTypeMetric': 'DES_BINARY',
           'matcherTypeSelector': 'SEL_NN',
           'detect_time_ms': 19,
           'desc_time_ms': 339,
           'match_time_ms': 0,
           'detect_keypoints_size': 1370,
           'match_keypoints_size': 0,
           'match_removed_keypoints_size': 0,
           'bVis': 0,
           'bLimitKpts': 0,
           'maxKeypoints': 50,
           
         },
         {
           'isError': '0',
           'image_name': '../images/KITTI/2011_09_26/image_00/data/0000000001.png',
           'detectorType': 'SHITOMASI',
           'descriptorType': 'BRISK',
           'matcherType': 'MAT_BF',
           'matcherTypeMetric': 'DES_BINARY',
           'matcherTypeSelector': 'SEL_NN',
           'detect_time_ms': 17,
           'desc_time_ms': 343,
           'match_time_ms': 0,
           'detect_keypoints_size': 1301,
           'match_keypoints_size': 125,
           'match_removed_keypoints_size': 0,
           'bVis': 0,
           'bLimitKpts': 0,
           'maxKeypoints': 50,
           
         },
         {
           'isError': '0',
           'image_name': '../images/KITTI/2011_09_26/image_00/data/0000000002.png',
           'detectorType': 'SHITOMASI',
           'descriptorType': 'BRISK',
           'matcherType': 'MAT_BF',
           'matcherTypeMetric': 'DES_BINARY',
           'matcherTypeSelector': 'SEL_NN',
           'detect_time_ms': 17,
           'desc_time_ms': 337,
           'match_time_ms': 0,
           'detect_keypoints_size': 1361,
           'match_keypoints_size': 118,
           'match_removed_keypoints_size': 0,
           'bVis': 0,
           'bLimitKpts': 0,
           'maxKeypoints': 50,
           
         },
         {
           'isError': '0',
           'image_name': '../images/KITTI/2011_09_26/image_00/data/0000000003.png',
           'detectorType': 'SHITOMASI',
           'descriptorType': 'BRISK',
           'matcherType': 'MAT_BF',
           'matcherTypeMetric': 'DES_BINARY',
           'matcherTypeSelector': 'SEL_NN',
           'detect_time_ms': 17,
           'desc_time_ms': 341,
           'match_time_ms': 0,
           'detect_keypoints_size': 1358,
           'match_keypoints_size': 123,
           'match_removed_keypoints_size': 0,
           'bVis': 0,
           'bLimitKpts': 0,
           'maxKeypoints': 50,
           
         },
         {
           'isError': '0',
           'image_name': '../images/KITTI/2011_09_26/image_00/data/0000000004.png',
           'detectorType': 'SHITOMASI',
           'descriptorType': 'BRISK',
           'matcherType': 'MAT_BF',
           'matcherTypeMetric': 'DES_BINARY',
           'matcherTypeSelector': 'SEL_NN',
           'detect_time_ms': 16,
           'desc_time_ms': 336,
           'match_time_ms': 0,
           'detect_keypoints_size': 1333,
           'match_keypoints_size': 120,
           'match_removed_keypoints_size': 0,
           'bVis': 0,
           'bLimitKpts': 0,
           'maxKeypoints': 50,
           
         },
         {
           'isError': '0',
           'image_name': '../images/KITTI/2011_09_26/image_00/data/0000000005.png',
           'detectorType': 'SHITOMASI',
           'descriptorType': 'BRISK',
           'matcherType': 'MAT_BF',
           'matcherTypeMetric': 'DES_BINARY',
           'matcherTypeSelector': 'SEL_NN',
           'detect_time_ms': 18,
           'desc_time_ms': 341,
           'match_time_ms': 0,
           'detect_keypoints_size': 1284,
           'match_keypoints_size': 120,
           'match_removed_keypoints_size': 0,
           'bVis': 0,
           'bLimitKpts': 0,
           'maxKeypoints': 50,
           
         },
         {
           'isError': '0',
           'image_name': '../images/KITTI/2011_09_26/image_00/data/0000000006.png',
           'detectorType': 'SHITOMASI',
           'descriptorType': 'BRISK',
           'matcherType': 'MAT_BF',
           'matcherTypeMetric': 'DES_BINARY',
           'matcherTypeSelector': 'SEL_NN',
           'detect_time_ms': 16,
           'desc_time_ms': 342,
           'match_time_ms': 0,
           'detect_keypoints_size': 1322,
           'match_keypoints_size': 113,
           'match_removed_keypoints_size': 0,
           'bVis': 0,
           'bLimitKpts': 0,
           'maxKeypoints': 50,
           
         },
         {
           'isError': '0',
           'image_name': '../images/KITTI/2011_09_26/image_00/data/0000000007.png',
           'detectorType': 'SHITOMASI',
           'descriptorType': 'BRISK',
           'matcherType': 'MAT_BF',
           'matcherTypeMetric': 'DES_BINARY',
           'matcherTypeSelector': 'SEL_NN',
           'detect_time_ms': 16,
           'desc_time_ms': 339,
           'match_time_ms': 0,
           'detect_keypoints_size': 1366,
           'match_keypoints_size': 114,
           'match_removed_keypoints_size': 0,
           'bVis': 0,
           'bLimitKpts': 0,
           'maxKeypoints': 50,
           
         },
         {
           'isError': '0',
           'image_name': '../images/KITTI/2011_09_26/image_00/data/0000000008.png',
           'detectorType': 'SHITOMASI',
           'descriptorType': 'BRISK',
           'matcherType': 'MAT_BF',
           'matcherTypeMetric': 'DES_BINARY',
           'matcherTypeSelector': 'SEL_NN',
           'detect_time_ms': 17,
           'desc_time_ms': 339,
           'match_time_ms': 0,
           'detect_keypoints_size': 1389,
           'match_keypoints_size': 123,
           'match_removed_keypoints_size': 0,
           'bVis': 0,
           'bLimitKpts': 0,
           'maxKeypoints': 50,
           
         },
        ]

results

results.csv

results.json

anylysis

report_analysis.md

report_analysis.html

import pandas as pd
import numpy as np
import os.path

def get_image_name(path): 
    return os.path.basename(path)

df = pd.read_csv("results.csv")
df['image_name'] = df['image_name'].apply(get_image_name)
del df['error']
df.head(3)

	image_name	detectorType	descriptorType	matcherType	matcherTypeMetric	matcherTypeSelector	detect_time	desc_time	detect_keypoints_size	match_keypoints_size	maxKeypoints
0	0000000000.png	SHITOMASI	BRISK	MAT_BF	DES_BINARY	SEL_NN	19	339	1370	0	50
1	0000000001.png	SHITOMASI	BRISK	MAT_BF	DES_BINARY	SEL_NN	17	343	1301	125	50
2	0000000002.png	SHITOMASI	BRISK	MAT_BF	DES_BINARY	SEL_NN	17	337	1361	118	50

detect_time by detectorType

df_detect= df.groupby(['detectorType']).mean()
df_detect.sort_values(by=['detect_time'], ascending=[1])

	detect_time	desc_time	match_time	detect_keypoints_size	match_keypoints_size	match_removed_keypoints_size	bVis	bLimitKpts	maxKeypoints
detectorType
FAST	0.933333	65.425000	0.739583	1787.4	84.781250	16.318750	0.0	0.0	50.0
ORB	7.445833	69.620833	0.427083	500.0	55.825000	12.200000	0.0	0.0	50.0
SHITOMASI	14.931250	65.900000	0.543750	1342.3	69.845833	10.179167	0.0	0.0	50.0
HARRIS	16.062500	64.222917	0.000000	173.7	13.500000	2.550000	0.0	0.0	50.0
AKAZE	72.608333	75.564583	0.972917	1342.9	121.081250	15.593750	0.0	0.0	50.0
SIFT	124.356250	71.014583	0.531250	1386.2	54.454167	18.220833	0.0	0.0	50.0
BRISK	365.677083	68.922917	1.629167	2711.6	145.668750	39.397917	0.0	0.0	50.0

detect_time by descriptorType

df_detect2= df.groupby(['descriptorType']).mean()
df_detect2.sort_values(by=['detect_time'], ascending=[1])

	detect_time	desc_time	match_time	detect_keypoints_size	match_keypoints_size	match_removed_keypoints_size	bVis	bLimitKpts	maxKeypoints
descriptorType
SIFT	82.450000	35.317857	1.037500	1320.585714	55.564286	8.078571	0.0	0.0	50.0
FREAK	85.591071	40.201786	0.776786	1320.585714	93.953571	23.675000	0.0	0.0	50.0
BRISK	86.833929	325.205357	0.864286	1320.585714	102.894643	23.191071	0.0	0.0	50.0
AKAZE	86.875000	8.635714	0.128571	1320.585714	19.337500	1.962500	0.0	0.0	50.0
BRIEF	86.998214	0.608929	0.667857	1320.585714	105.757143	21.528571	0.0	0.0	50.0
ORB	87.264286	2.033929	0.676786	1320.585714	89.769643	19.673214	0.0	0.0	50.0

desc_time by detectorType

df_detect= df.groupby(['detectorType']).mean()
df_detect.sort_values(by=['desc_time'], ascending=[1])

	detect_time	desc_time	match_time	detect_keypoints_size	match_keypoints_size	match_removed_keypoints_size	bVis	bLimitKpts	maxKeypoints
detectorType
HARRIS	16.062500	64.222917	0.000000	173.7	13.500000	2.550000	0.0	0.0	50.0
FAST	0.933333	65.425000	0.739583	1787.4	84.781250	16.318750	0.0	0.0	50.0
SHITOMASI	14.931250	65.900000	0.543750	1342.3	69.845833	10.179167	0.0	0.0	50.0
BRISK	365.677083	68.922917	1.629167	2711.6	145.668750	39.397917	0.0	0.0	50.0
ORB	7.445833	69.620833	0.427083	500.0	55.825000	12.200000	0.0	0.0	50.0
SIFT	124.356250	71.014583	0.531250	1386.2	54.454167	18.220833	0.0	0.0	50.0
AKAZE	72.608333	75.564583	0.972917	1342.9	121.081250	15.593750	0.0	0.0	50.0

desc_time by descriptorType

df_detect= df.groupby(['descriptorType']).mean()
df_detect.sort_values(by=['desc_time'], ascending=[1])

	detect_time	desc_time	match_time	detect_keypoints_size	match_keypoints_size	match_removed_keypoints_size	bVis	bLimitKpts	maxKeypoints
descriptorType
BRIEF	86.998214	0.608929	0.667857	1320.585714	105.757143	21.528571	0.0	0.0	50.0
ORB	87.264286	2.033929	0.676786	1320.585714	89.769643	19.673214	0.0	0.0	50.0
AKAZE	86.875000	8.635714	0.128571	1320.585714	19.337500	1.962500	0.0	0.0	50.0
SIFT	82.450000	35.317857	1.037500	1320.585714	55.564286	8.078571	0.0	0.0	50.0
FREAK	85.591071	40.201786	0.776786	1320.585714	93.953571	23.675000	0.0	0.0	50.0
BRISK	86.833929	325.205357	0.864286	1320.585714	102.894643	23.191071	0.0	0.0	50.0

detect_time by descriptorType, detectorType

df_detect3= df.groupby(['descriptorType', 'detectorType']).mean()
df_detect3.sort_values(by=['detect_time'], ascending=[0])

		detect_time	desc_time	match_time	detect_keypoints_size	match_keypoints_size	match_removed_keypoints_size	bVis	bLimitKpts	maxKeypoints
descriptorType	detectorType
ORB	BRISK	369.7000	4.0750	1.8500	2711.6	187.4500	63.3500	0.0	0.0	50.0
BRIEF	BRISK	367.4750	1.0125	1.6750	2711.6	202.2250	48.5750	0.0	0.0	50.0
BRISK	BRISK	366.1625	327.3625	2.1500	2711.6	197.6250	53.1750	0.0	0.0	50.0
AKAZE	BRISK	364.1625	0.0000	0.0000	2711.6	0.0000	0.0000	0.0	0.0	50.0
FREAK	BRISK	363.5875	40.2000	1.8875	2711.6	182.4250	50.1750	0.0	0.0	50.0
SIFT	BRISK	362.9750	40.8875	2.2125	2711.6	104.2875	21.1125	0.0	0.0	50.0
ORB	SIFT	128.9125	0.0000	0.0000	1386.2	0.0000	0.0000	0.0	0.0	50.0
FREAK	SIFT	128.7250	39.4625	0.7875	1386.2	89.5625	34.3375	0.0	0.0	50.0
AKAZE	SIFT	127.7625	0.0000	0.0000	1386.2	0.0000	0.0000	0.0	0.0	50.0
BRIEF	SIFT	127.7500	0.5500	0.5750	1386.2	95.1125	29.7875	0.0	0.0	50.0
BRISK	SIFT	127.0875	305.2000	0.8750	1386.2	90.7375	34.0625	0.0	0.0	50.0
SIFT	SIFT	105.9000	80.8750	0.9500	1386.2	51.3125	11.1375	0.0	0.0	50.0
BRISK	AKAZE	73.9875	325.3750	0.9500	1342.9	132.8250	16.2750	0.0	0.0	50.0
AKAZE	AKAZE	73.3750	60.4500	0.9000	1342.9	135.3625	13.7375	0.0	0.0	50.0
ORB	AKAZE	73.0750	3.1875	0.9625	1342.9	127.0250	22.0750	0.0	0.0	50.0
BRIEF	AKAZE	72.7750	0.9500	0.8000	1342.9	133.4750	15.6250	0.0	0.0	50.0
FREAK	AKAZE	72.2250	39.6500	0.9250	1342.9	128.5750	20.5250	0.0	0.0	50.0
SIFT	AKAZE	70.2125	23.7750	1.3000	1342.9	69.2250	5.3250	0.0	0.0	50.0
AKAZE	HARRIS	17.8000	0.0000	0.0000	173.7	0.0000	0.0000	0.0	0.0	50.0
BRISK	SHITOMASI	16.9000	336.1125	0.6625	1342.3	89.7375	16.9625	0.0	0.0	50.0
SIFT	HARRIS	16.7500	15.7500	0.0000	173.7	9.4250	1.2750	0.0	0.0	50.0
BRIEF	HARRIS	16.7000	0.0750	0.0000	173.7	18.5500	2.8500	0.0	0.0	50.0
BRIEF	SHITOMASI	16.2125	1.0125	0.5000	1342.3	97.3000	9.4000	0.0	0.0	50.0
ORB	SHITOMASI	16.1375	1.0000	0.5750	1342.3	95.2750	11.4250	0.0	0.0	50.0
AKAZE	SHITOMASI	15.8750	0.0000	0.0000	1342.3	0.0000	0.0000	0.0	0.0	50.0
BRISK	HARRIS	15.4750	327.7625	0.0000	173.7	17.2750	4.1250	0.0	0.0	50.0
ORB	HARRIS	15.0125	0.6500	0.0000	173.7	18.3750	3.0250	0.0	0.0	50.0
FREAK	HARRIS	14.6375	41.1000	0.0000	173.7	17.3750	4.0250	0.0	0.0	50.0
SIFT	SHITOMASI	12.7250	16.0125	0.9125	1342.3	49.8500	3.5000	0.0	0.0	50.0
FREAK	SHITOMASI	11.7375	41.2625	0.6125	1342.3	86.9125	19.7875	0.0	0.0	50.0
AKAZE	ORB	8.1750	0.0000	0.0000	500.0	0.0000	0.0000	0.0	0.0	50.0
SIFT	ORB	7.5875	47.6750	0.7375	500.0	44.9750	6.6750	0.0	0.0	50.0
FREAK	ORB	7.2625	39.0625	0.3250	500.0	46.6250	8.2750	0.0	0.0	50.0
BRIEF	ORB	7.2375	0.1875	0.4500	500.0	76.5625	26.7375	0.0	0.0	50.0
BRISK	ORB	7.2250	326.4750	0.5125	500.0	82.5750	12.4250	0.0	0.0	50.0
ORB	ORB	7.1875	4.3250	0.5375	500.0	84.2125	19.0875	0.0	0.0	50.0
BRISK	FAST	1.0000	328.1500	0.9000	1787.4	109.4875	25.3125	0.0	0.0	50.0
SIFT	FAST	1.0000	22.2500	1.1500	1787.4	59.8750	7.5250	0.0	0.0	50.0
AKAZE	FAST	0.9750	0.0000	0.0000	1787.4	0.0000	0.0000	0.0	0.0	50.0
FREAK	FAST	0.9625	40.6750	0.9000	1787.4	106.2000	28.6000	0.0	0.0	50.0
BRIEF	FAST	0.8375	0.4750	0.6750	1787.4	117.0750	17.7250	0.0	0.0	50.0
ORB	FAST	0.8250	1.0000	0.8125	1787.4	116.0500	18.7500	0.0	0.0	50.0

desc_time by descriptorType, detectorType

df_detect3= df.groupby(['descriptorType', 'detectorType']).mean()
df_detect3.sort_values(by=['desc_time'], ascending=[0])

		detect_time	desc_time	match_time	detect_keypoints_size	match_keypoints_size	match_removed_keypoints_size	bVis	bLimitKpts	maxKeypoints
descriptorType	detectorType
BRISK	SHITOMASI	16.9000	336.1125	0.6625	1342.3	89.7375	16.9625	0.0	0.0	50.0
	FAST	1.0000	328.1500	0.9000	1787.4	109.4875	25.3125	0.0	0.0	50.0
	HARRIS	15.4750	327.7625	0.0000	173.7	17.2750	4.1250	0.0	0.0	50.0
	BRISK	366.1625	327.3625	2.1500	2711.6	197.6250	53.1750	0.0	0.0	50.0
	ORB	7.2250	326.4750	0.5125	500.0	82.5750	12.4250	0.0	0.0	50.0
	AKAZE	73.9875	325.3750	0.9500	1342.9	132.8250	16.2750	0.0	0.0	50.0
	SIFT	127.0875	305.2000	0.8750	1386.2	90.7375	34.0625	0.0	0.0	50.0
SIFT	SIFT	105.9000	80.8750	0.9500	1386.2	51.3125	11.1375	0.0	0.0	50.0
AKAZE	AKAZE	73.3750	60.4500	0.9000	1342.9	135.3625	13.7375	0.0	0.0	50.0
SIFT	ORB	7.5875	47.6750	0.7375	500.0	44.9750	6.6750	0.0	0.0	50.0
FREAK	SHITOMASI	11.7375	41.2625	0.6125	1342.3	86.9125	19.7875	0.0	0.0	50.0
FREAK	HARRIS	14.6375	41.1000	0.0000	173.7	17.3750	4.0250	0.0	0.0	50.0
SIFT	BRISK	362.9750	40.8875	2.2125	2711.6	104.2875	21.1125	0.0	0.0	50.0
FREAK	FAST	0.9625	40.6750	0.9000	1787.4	106.2000	28.6000	0.0	0.0	50.0
	BRISK	363.5875	40.2000	1.8875	2711.6	182.4250	50.1750	0.0	0.0	50.0
	AKAZE	72.2250	39.6500	0.9250	1342.9	128.5750	20.5250	0.0	0.0	50.0
	SIFT	128.7250	39.4625	0.7875	1386.2	89.5625	34.3375	0.0	0.0	50.0
	ORB	7.2625	39.0625	0.3250	500.0	46.6250	8.2750	0.0	0.0	50.0
SIFT	AKAZE	70.2125	23.7750	1.3000	1342.9	69.2250	5.3250	0.0	0.0	50.0
	FAST	1.0000	22.2500	1.1500	1787.4	59.8750	7.5250	0.0	0.0	50.0
	SHITOMASI	12.7250	16.0125	0.9125	1342.3	49.8500	3.5000	0.0	0.0	50.0
	HARRIS	16.7500	15.7500	0.0000	173.7	9.4250	1.2750	0.0	0.0	50.0
ORB	ORB	7.1875	4.3250	0.5375	500.0	84.2125	19.0875	0.0	0.0	50.0
	BRISK	369.7000	4.0750	1.8500	2711.6	187.4500	63.3500	0.0	0.0	50.0
	AKAZE	73.0750	3.1875	0.9625	1342.9	127.0250	22.0750	0.0	0.0	50.0
BRIEF	SHITOMASI	16.2125	1.0125	0.5000	1342.3	97.3000	9.4000	0.0	0.0	50.0
BRIEF	BRISK	367.4750	1.0125	1.6750	2711.6	202.2250	48.5750	0.0	0.0	50.0
ORB	FAST	0.8250	1.0000	0.8125	1787.4	116.0500	18.7500	0.0	0.0	50.0
ORB	SHITOMASI	16.1375	1.0000	0.5750	1342.3	95.2750	11.4250	0.0	0.0	50.0
BRIEF	AKAZE	72.7750	0.9500	0.8000	1342.9	133.4750	15.6250	0.0	0.0	50.0
ORB	HARRIS	15.0125	0.6500	0.0000	173.7	18.3750	3.0250	0.0	0.0	50.0
BRIEF	SIFT	127.7500	0.5500	0.5750	1386.2	95.1125	29.7875	0.0	0.0	50.0
	FAST	0.8375	0.4750	0.6750	1787.4	117.0750	17.7250	0.0	0.0	50.0
	ORB	7.2375	0.1875	0.4500	500.0	76.5625	26.7375	0.0	0.0	50.0
	HARRIS	16.7000	0.0750	0.0000	173.7	18.5500	2.8500	0.0	0.0	50.0
AKAZE	HARRIS	17.8000	0.0000	0.0000	173.7	0.0000	0.0000	0.0	0.0	50.0
ORB	SIFT	128.9125	0.0000	0.0000	1386.2	0.0000	0.0000	0.0	0.0	50.0
AKAZE	SIFT	127.7625	0.0000	0.0000	1386.2	0.0000	0.0000	0.0	0.0	50.0
	SHITOMASI	15.8750	0.0000	0.0000	1342.3	0.0000	0.0000	0.0	0.0	50.0
	BRISK	364.1625	0.0000	0.0000	2711.6	0.0000	0.0000	0.0	0.0	50.0
	FAST	0.9750	0.0000	0.0000	1787.4	0.0000	0.0000	0.0	0.0	50.0
	ORB	8.1750	0.0000	0.0000	500.0	0.0000	0.0000	0.0	0.0	50.0

Name		Name	Last commit message	Last commit date
Latest commit History 42 Commits
images		images
src		src
.gitignore		.gitignore
CMakeLists.txt		CMakeLists.txt
LICENSE		LICENSE
README.md		README.md
report_analysis.html		report_analysis.html
report_analysis.ipynb		report_analysis.ipynb
report_analysis.md		report_analysis.md
results.csv		results.csv
results.json		results.json
runtest.log		runtest.log

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

SFND 2D Feature Tracking

Dependencies for Running Locally

Basic Build Instructions

Code

Benchmark

audit log for m looping multiple options and collect experiment data

results

anylysis

About

Releases

Packages

Languages

License

techprav2010/SFND_2D_Feature_Tracking

Folders and files

Latest commit

History

Repository files navigation

SFND 2D Feature Tracking

Dependencies for Running Locally

Basic Build Instructions

Code

Benchmark

audit log for m looping multiple options and collect experiment data

results

anylysis

About

Resources

License

Stars

Watchers

Forks

Releases

Packages 0

Languages

Packages