Skip to content

Latest commit

 

History

History
103 lines (72 loc) · 4.33 KB

train-usage.md

File metadata and controls

103 lines (72 loc) · 4.33 KB

Keras RetinaNet on Liza Alert Drone Data Set Build Status

Keras implementation of RetinaNet object detection as described in Focal Loss for Dense Object Detection.

This code is borrowed from Keras Implementation of this model at https://github.com/fizyr/keras-retinanet and updated to run on Liza Aletr Drone Dataset

Installation

  1. Clone this repository.
  2. cd RetinaNet
  3. Ensure numpy is installed using pip install numpy --user
  4. In the repository, execute pip install . --user. Note that due to inconsistencies with how tensorflow should be installed, this package does not define a dependency on tensorflow as it will try to install that (which at least on Arch Linux results in an incorrect installation). Please make sure tensorflow is installed as per your systems requirements.
  5. Alternatively, you can run the code directly from the cloned repository, however you need to run python setup.py build_ext --inplace to compile Cython code first.
  6. Optionally, install pycocotools if you want to train / test on the MS COCO dataset by running pip install --user git+https://github.com/cocodataset/cocoapi.git#subdirectory=PythonAPI.

Training on custom data set (added by Priyanka Dwivedi, Georgy Perevozchikov)

For training on a [custom dataset], a CSV file can be used as a way to pass the data. See below for more details on the format of these CSV files.

To train using your CSV, run:

keras_retinanet/bin/train.py --weights snapshots/resnet50_coco_best_v2.1.0.h5 csv ../data/SDD-CSV/train_annotations.csv labels.csv --val-annotations ../data/SDD-CSV/val_annotations.csv

First step to train on SDD dataset

To train using your Pascal VOC, run:

keras_retinanet/bin/train.py --weights snapshots/resnet50_base_best.h5 --freeze-backbone --config config.ini pascal ../data/LADD

Second step to train on LADD dataset and fine-tuning model

Running directly from RetinaNet folder in the repository

Here

  • weights: Path to the weights for initializing training
  • csv indicates retinanet is trained on a custom data set
  • train_annotations.csv is path to training annotations
  • labels.csv are the labels in the format class_name, class_id with 0 reserved for background class
  • val_annotations is path to validation annotations
  • config.ini configuration file

If your dataset is small, you can also use the --freeze-backbone argument to freeze the backbone layers.

Training Set

I uploaded the images used for training and validation to the yandex disk link below. Please download the same: https://yadi.sk/d/4Hz_1qpiNbHhpQ

Annotations format

The CSV file with annotations should contain one annotation per line. Images with multiple bounding boxes should use one row per bounding box. Note that indexing for pixel values starts at 0. The expected format of each line is:

path/to/image.jpg,x1,y1,x2,y2,class_name

Labels format

The class name to ID mapping file should contain one mapping per line. Each line should use the following format:

class_name,id

Evaluating Results (added by Priyanka Dwivedi and Georgy Perevozchikov)

To calculate mean average precision on the validation set, please run

keras_retinanet/bin/evaluate.py csv val_annotations.csv labels.csv snapshots/resnet50_csv_31_interface.h5

or for Pascal VOC

keras_retinanet/bin/evaluate.py pascal ../data/LADD snapshots/resnet50_pascal_06.h5  --convert-model

Here we pass the val_annotations, labels and path to the trained weights

Running Inference on Images and Videos (added by Priyanka Dwivedi and Georgy Perevozchikov)

To run inference on the trained model, first step is to convert the trained model to a format that can be used by inference. The command for this is:

keras_retinanet/bin/convert_model.py snapshots/resnet50_pascal_08.h5 snapshots/resnet50_pascal_08_inference.h5

Here first path is the path to the trained model and the second would be the path to the converted inference model

I created two notebooks from the original code that can be used to run inference on images and on videos.

The notebooks are:

  • RetinaNet/ResNet50RetinaNet-Image.ipynb : For inference on a batch of images
  • RetinaNet/ResNet50RetinaNet-Video.ipynb : For inference on a video