OpenISS keras-yolo3

This repo is a clone from qqwweee/keras-yolo3 hosted on GitHub, with modification to adjust to person re-identification model only (from YOLOv3 object detection model). The prediction code is obtained from this repo and adapted to current implementation. The result figure is generated by using the code ported from this repo.

Introduction

A Keras implementation of YOLOv3 (Tensorflow backend) inspired by allanzelener/YAD2K.

This is a part of the Eric Lai's ML portion of the OpenISS project for his master's thesis:

Haotao Lai, An OpenISS Framework Specialization for Person Re-identification, Master's thesis, August 2019, Concordia University, Montreal

Environment

Images and videos can be from any source, but we provided you with a sample of video and images in folders video and image. This demo can be run in an interactive and non-interactive way for both image and video default is non-interactive. You can enable interactive by passing --interactive parameters.

For image you provide the server with image path/name and return the classes of images.
For video you will provide video path/name in the script argument. The output for interactive mode is a live video that dispaly classified object in each video frame. The new classified video will be stored in your video folder for both interactive and non-interactive modes.

The video sample is from youtube channel Watched Walker.

Images are from YAD2K repostry project.

Quick Start

Download YOLOv3 weights from YOLO website.
Convert the Darknet YOLO model to a Keras model.
Run YOLO detection.

wget https://pjreddie.com/media/files/yolov3.weights
python convert.py yolov3.cfg yolov3.weights model_data/yolo.h5
python yolo_video.py [OPTIONS...] --image, for image detection mode, OR
python yolo_video.py [video_path] [output_path (optional)]

For image run follwing command and the program will ask you about the image path and name:

python yolo_video.py --model model_data/yolo.h5 --classes model_data/coco_classes.txt --image

For video non-interactive default

python yolo_video.py --input video/v1.avi --output vido/001.avi

For video interactive

python yolo_video.py --input video/v1.avi --output vido/001.avi  --interactive

For Tiny YOLOv3, just do in a similar way, just specify model path and anchor path with --model model_file and --anchors anchor_file.

Usage

Use --help to see usage of yolo_video.py:

usage: yolo_video.py [-h] [--model MODEL] [--anchors ANCHORS]
                     [--classes CLASSES] [--gpu_num GPU_NUM] [--image]
                     [--input] [--output] [--interactive]

positional arguments:
  --input        Video input path
  --output       Video output path

optional arguments:
  -h, --help         show this help message and exit
  --model MODEL      path to model weight file, default model_data/yolo.h5
  --anchors ANCHORS  path to anchor definitions, default
                     model_data/yolo_anchors.txt
  --classes CLASSES  path to class definitions, default
                     model_data/coco_classes.txt
  --gpu_num GPU_NUM  Number of GPU to use, default 1
  --image            Image detection mode, will ignore all positional arguments
  --interactive      To show live video classification

MultiGPU usage: use --gpu_num N to use N GPUs. It is passed to the Keras multi_gpu_model().

Training

Generate your own annotation file and class names file. One row for one image; Row format: image_file_path box1 box2 ... boxN; Box format: x_min,y_min,x_max,y_max,class_id (no space). For VOC dataset, try python voc_annotation.py Here is an example:
```
path/to/img1.jpg 50,100,150,200,0 30,50,200,120,3
path/to/img2.jpg 120,300,250,600,2
...
```
Make sure you have run python convert.py -w yolov3.cfg yolov3.weights model_data/yolo_weights.h5 The file model_data/yolo_weights.h5 is used to load pretrained weights.
Modify train.py and start training. python train.py Use your trained weights or checkpoint weights with command line option --model model_file when using yolo_video.py Remember to modify class path or anchor path, with --classes class_file and --anchors anchor_file.

If you want to use original pretrained weights for YOLOv3: 1. wget https://pjreddie.com/media/files/darknet53.conv.74 2. rename it as darknet53.weights 3. python convert.py -w darknet53.cfg darknet53.weights model_data/darknet53_weights.h5 4. use model_data/darknet53_weights.h5 in train.py

Some issues to know

The test environment is
- Python 3.5.2
- Keras 2.1.5
- tensorflow-gpu 1.10.0
Default anchors are used. If you use your own anchors, probably some changes are needed.
The inference result is not totally the same as Darknet but the difference is small.
The speed is slower than Darknet. Replacing PIL with opencv may help a little.
Always load pretrained weights and freeze layers in the first stage of training. Or try Darknet training. It's OK if there is a mismatch warning.
The training strategy is for reference only. Adjust it according to your dataset and your goal. And add further strategy if needed.
For speeding up the training process with frozen layers train_bottleneck.py can be used. It will compute the bottleneck features of the frozen model first and then only trains the last layers. This makes training on CPU possible in a reasonable time. See this for more information on bottleneck features.