siddharthganesh / dpm-cfnet

SiamFC - TensorFlow

TensorFlow port of the tracking method described in the paper Fully-Convolutional Siamese nets for object tracking.

In particular, it is the improved version presented as baseline in End-to-end representation learning for Correlation Filter based tracking, which achieves state-of-the-art performance at high framerate. The other methods presented in the paper (similar performance, shallower network) haven't been ported yet.

Settings things up with virtualenv

1. Get virtualenv if you don't have it already pip install virtualenv
2. Create new virtualenv with Python 2.7 virtualenv --python=/usr/bin/python2.7 ve-tracking
3. Activate the virtualenv source ~/tracking-ve/bin/activate
4. Clone the repository git clone https://github.com/torrvision/siamfc-tf.git
5. cd siamfc-tf
6. Install the required packages sudo pip install -r requirements.txt
7. mkdir pretrained data
8. Download the pretrained networks in pretrained and unzip the archive (we will only use baseline-conv5_e55.mat)
9. Download video sequences in data and unzip the archive.

Running the tracker

1. Set video from parameters.evaluation to "all" or to a specific sequence (e.g. "vot2016_ball1")
2. See if you are happy with the default parameters in parameters/hyperparameters.json
3. Optionally enable visualization in parameters/run.json
4. Call the main script (within an active virtualenv session) python run_tracker_evaluation.py

dpm-cfnet's Issues

i run run_tracker_evaluation.py and i can not understand what is that output like Precisions AUC ,Precisions means ,can you explain for me and i wil appricate it .

and i also doubt this script can not use the GPU effcecently,because i just have a result after 10 hours ,can you help me to figure it out ,thank you so much.

CUDA_VISIBLE_DEVICES='0,1' python run.py
Using Tensorflow 1.1.0

Layer 1
CONV: setting br_conv1f br_conv1b
CONV: stride 2, filter-group False
BNORM: setting br_bn1b br_bn1m br_bn1x
MAX-POOL: size 3 and stride 2
Layer 2
CONV: setting br_conv2f br_conv2b
CONV: stride 1, filter-group True
BNORM: setting br_bn2b br_bn2m br_bn2x
MAX-POOL: size 3 and stride 1
Layer 3
CONV: setting br_conv3f br_conv3b
CONV: stride 1, filter-group False
BNORM: setting br_bn3b br_bn3m br_bn3x
Layer 4
CONV: setting br_conv4f br_conv4b
CONV: stride 1, filter-group True
BNORM: setting br_bn4b br_bn4m br_bn4x
Layer 5
CONV: setting br_conv5f br_conv5b
CONV: stride 1, filter-group True

0 -- tc_Airport_ce -- Precision: 10.81 -- Precisions AUC: 3.44 -- IOU: 12.57 -- Speed: 0.20 --

1 -- tc_Baby_ce -- Precision: 96.62 -- Precisions AUC: 38.98 -- IOU: 62.22 -- Speed: 0.19 --

2 -- tc_Badminton_ce1 -- Precision: 32.12 -- Precisions AUC: 13.52 -- IOU: 23.51 -- Speed: 0.19 --

i also want to know if i can train the cfnet to get my own model as the type of tensorflow? because i did not fingure the construct of your amazing work ,look forward you message and thank you so much