https://github.com/vinay0410/pedestrian_detection

Detects Pedestrians in images using HOG as a feature extractor and SVM for classification
https://github.com/vinay0410/pedestrian_detection

computer-vision detects-pedestrians hog-features human-detection human-detection-algorithm pedestrian-detection person-detection svm

Last synced: 24 days ago
JSON representation

Detects Pedestrians in images using HOG as a feature extractor and SVM for classification

• Host: GitHub
• URL: https://github.com/vinay0410/pedestrian_detection
• Owner: vinay0410
• License: gpl-3.0
• Created: 2017-08-28T14:30:34.000Z (over 7 years ago)
• Default Branch: master
• Last Pushed: 2022-10-18T13:59:28.000Z (over 2 years ago)
• Last Synced: 2025-03-26T22:51:18.371Z (about 1 month ago)
• Topics: computer-vision, detects-pedestrians, hog-features, human-detection, human-detection-algorithm, pedestrian-detection, person-detection, svm
• Language: Python
• Homepage:
• Size: 2.61 MB
• Stars: 74
• Watchers: 5
• Forks: 25
• Open Issues: 5
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Pedestrian Detection

**Note:** The code provided may not be compatible with python3. A migration to python3 is in the pipeline.

## Dependencies

* OpenCV

* scikit-image ```pip install scikit-image==0.14.1```

* scikkit-learn ```pip install scikit-learn==0.20.2```

## Running Detection

To test on images, simply run, `python detectmulti.py -i `

For example, `python detectmulti.py -i sample_images/pedestrian.jpg`

For more options run, `python detectmulti.py -h`

Following are some examples,

![Pedestrian](.readme_images/before_nms1.png?raw=true "Sample Results")

![Pedestrian](.readme_images/after_nms1.png?raw=true "Sample Results")

![Pedestrian](.readme_images/before_nms2.png?raw=true "Sample Results")

![Pedestrian](.readme_images/after_nms2.png?raw=true "Sample Results")

## Training

This code is meant for Training a Pedestrian Detector using [INRIA Person Dataset](http://pascal.inrialpes.fr/data/human/).  

**Note:** If the above link doesn't work, then the above dataset can be downloaded from [here](https://drive.google.com/file/d/14GD_pBpBsprPiZlkmtXN_y5K72To16if/view?usp=sharing).

Download, Uncompress and place it in the root of this repository.

Then just run:

```

sudo sh fixpng.sh # To fix the broken png files in this dataset.

```

**Note:** *The above step is necessary to fix the broken pngs in the dataset, without running this the classifer won't be correctly trained.*

To train just run:

```

python train.py --pos  --neg 

```

For INRIA dataset, this would be probably,

```

python train.py --pos INRIAPerson/train_64x128_H96/pos --neg INRIAPerson/train_64x128_H96/neg

```

After training, two new files would be created, namely, `person.pkl` and `person_final.pkl`. The former is the pre eliminary detector and the latter is the improved (hard negatively mined) detector.

### Note on Training:

Training can use high amounts of memory, so be sure to have a swap space in case of RAM overflows. Also, memory consumption can be reduced by decreasing the maximum number of hard negative windows to be mined. This is defined by the `MAX_HARD_NEGATIVES` global variable in `train.py`.

## Testing

After successful training just run:

```

python test.py --pos  --neg 

```

For INRIA dataset, this would be probably,

```

python test.py --pos INRIAPerson/test_64x128_H96/pos --neg INRIAPerson/test_64x128_H96/neg

```

This would print `True Positives`, `True Negatives`, `False Positives`, `False Negatives`, `Precision`, `Recall` and `F1 Score`.