https://github.com/kiritigowda/mivisionx-classifier
Live Image Classification
https://github.com/kiritigowda/mivisionx-classifier
classification concovultion googlenet image inceptionv4 net neural resnet-101 resnet-152 resnet-50 vgg16 vgg19
Last synced: 30 days ago
JSON representation
Live Image Classification
- Host: GitHub
- URL: https://github.com/kiritigowda/mivisionx-classifier
- Owner: kiritigowda
- License: mit
- Created: 2018-11-12T17:31:12.000Z (almost 7 years ago)
- Default Branch: master
- Last Pushed: 2019-04-03T20:57:52.000Z (over 6 years ago)
- Last Synced: 2025-01-14T07:15:13.164Z (9 months ago)
- Topics: classification, concovultion, googlenet, image, inceptionv4, net, neural, resnet-101, resnet-152, resnet-50, vgg16, vgg19
- Language: C++
- Homepage: https://kiritigowda.github.io/MIVisionX-Classifier/
- Size: 763 KB
- Stars: 1
- Watchers: 2
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
[](https://opensource.org/licenses/MIT)
[](https://travis-ci.org/kiritigowda/MIVisionX-Classifier)
# Live Image Classification
This application runs know CNN image classifiers on live/pre-recorded video stream.
## MIVisionX Image Classification Control
## MIVisionX Image Classification
## Usage
### Prerequisites
Build & Install [MIVisionX](https://github.com/GPUOpen-ProfessionalCompute-Libraries/MIVisionX#build--install-mivisionx)
### Build
````
git clone https://github.com/kiritigowda/MIVisionX-Classifier
cd MIVisionX-Classifier
cmake .
make
````
### Run
```
Usage: ./classifier
--inception [optional]
--resnet50 [optional]
--vgg16 [optional]
--googlenet [optional]
--resnet101 [optional]
--resnet152 [optional]
--vgg19 [optional]
--label [required]
--video /<--capture 0> [required]
```
#### Generating weights.bin for different Models
Download or train your own caffemodel and run the [model_compiler](https://github.com/GPUOpen-ProfessionalCompute-Libraries/MIVisionX/tree/master/model_compiler#neural-net-model-compiler--optimizer) to get the corresponding weights.bin file.
1. Download or train your own caffemodel.
Here is the sample download link that contains all the prototxt:
https://github.com/SnailTyan/caffe-model-zoo
2. Using [Model Compiler](https://github.com/GPUOpen-ProfessionalCompute-Libraries/MIVisionX/tree/master/model_compiler#neural-net-model-compiler--optimizer)
To convert a caffemodel into AMD NNIR model:
```
% python caffe_to_nnir.py --input-dims [--verbose <0|1>]
```
To convert an AMD NNIR model into OpenVX C code:
````
% python nnir_to_openvx.py
````
The weights file will be generated and you can use that as an input for this project.
#### --label text
The labels.txt file in this project.
#### --video file
Test the classification on your own video. Give the path to your video.
#### --capture 0
If you want to test with a live cam, turn on this option.
### Example
```
./classifier --inception /PATH/TO/inceptionV4/weights.bin --resnet50/PATH/TO/resnet50/weights.bin ...
...
--vgg19 /PATH/TO/vgg19/weights.bin
--label labels.txt
--capture 0
```