This demo demonstrates how to run Image Retrieval models using OpenVINO™.
NOTE: Only batch size of 1 is supported.
How It Works
The demo application expects an image retrieval model in the Intermediate Representation (IR) format.
As input, the demo application takes:
- a path to a list of images represented by textfile with following format 'path_to_image' 'ID'
- a path to a video file or a device node of a web-camera specified with a command line argument
The demo workflow is the following:
- The demo application reads video frames one by one, runs ROI detector that extracts ROI (moving area).
- Extracted ROI is passed to artificial neural network that computes embedding vector for extracted frame area.
- Then the demo application searches computed embedding in gallery of images in order to determine which image in the gallery is the most similar to what one can see on video frame.
- The app visualizes results of it work as graphical window where following objects are shown.
- Input frame with detected ROI.
- Top-10 most similar images from the gallery.
- Performance characteristics.
NOTE: By default, Open Model Zoo demos expect input with BGR channels order. If you trained your model to work with RGB order, you need to manually rearrange the default channels order in the demo application or reconvert your model using the Model Optimizer tool with
--reverse_input_channels argument specified. For more information about the argument, refer to When to Reverse Input Channels section of Converting a Model Using General Conversion Parameters.
Run the application with the
-h option to see the following usage message:
usage: image_retrieval_demo.py [-h] -m MODEL -i INPUT [--loop]
[-o OUTPUT] [-limit OUTPUT_LIMIT]
-g GALLERY [-gt GROUND_TRUTH]
[-d DEVICE] [-l CPU_EXTENSION]
[--no_show] [-u UTILIZATION_MONITORS]
-h, --help Show this help message and exit.
-m MODEL, --model MODEL
Required. Path to an .xml file with a trained model.
-i INPUT, --input INPUT
Required. Path to a video file or a device node of a
--loop Optional. Enable reading the input in a loop.
-o OUTPUT, --output OUTPUT
Optional. Name of output to save.
-limit OUTPUT_LIMIT, --output_limit OUTPUT_LIMIT
Optional. Number of frames to store in output.
If 0 is set, all frames are stored.
-g GALLERY, --gallery GALLERY
Required. Path to a file listing gallery images.
-gt GROUND_TRUTH, --ground_truth GROUND_TRUTH
Optional. Ground truth class.
-d DEVICE, --device DEVICE
Optional. Specify the target device to infer on: CPU,
GPU, FPGA, HDDL or MYRIAD. The demo will look for a
suitable plugin for device specified (by default, it
-l CPU_EXTENSION, --cpu_extension CPU_EXTENSION
Optional. Required for CPU custom layers. Absolute
path to a shared library with the kernels
--no_show Optional. Do not visualize inference results.
-u UTILIZATION_MONITORS, --utilization_monitors UTILIZATION_MONITORS
Optional. List of monitors to show initially.
Running the application with an empty list of options yields the short version of the usage message and an error message.
To run the demo, you can use public or pre-trained models. To download the pre-trained models, use the OpenVINO Model Downloader. The list of models supported by the demo is in the
models.lst file in the demo's directory.
NOTE: Before running the demo with a trained model, make sure the model is converted to the Inference Engine format (
*.bin) using the Model Optimizer tool.
To run the demo, please provide paths to the model in the IR format, to a file with class labels, and to an input video, image, or folder with images:
python image_retrieval_demo.py \
-m <path_to_model>/image-retrieval-0001.xml \
-i <path_to_video>/video.dav.mp4 \
-g <path_to_file>/list.txt \
An example of file listing gallery images can be found here.
Examples of videos can be found here.
The application uses OpenCV to display gallery searching result and current inference performance.