|| Testing custom Object detector with tensorflow object detection api ||
| Vision of this video |
Detect cat and dog along with bounding box using tensorflow object detection api
| What is special in this video |
If the image is grayscale, model works and show the prediction along with the bounding box.
| Steps implemented while creating the model |
1.) Created the annotation for cat and dog images.
2.) Converted the xml file into the csv file
3.) Converted the csv file into tf record
4.) Trained the images with the pretrained model. I have used ssd mobilenet. Please select the model from here
https://github.com/tensorflow/models/...
5.) Created the customized model from the checkpoints created in the step 4. Trained model contains the frozen file which is used for the prediction.
6.) Real time testing with the new model which I am demonstrating in the video.
| Some Suggestions |
I would like you to create the virtualenv if any one wants to train his or her model. To create the virtualenv in ubuntu, please follow the below command
virtualenv --python python3 *name_env*
You will also need requirements.txt file to install the required packages. Please do find the file from here (https://bit.ly/2MiQ2Lq). Please note that this video is not for training the model but for testing the custom model. I have named the new model as dog_cat_inference_graph as it detects cat and dog.
| Some references and resources |
http://www.aisangam.com/blog/tensorfl...
https://github.com/AISangam
| For queries|
Email: [email protected]
Skype: aisangamofficial
| Join us |
https://www.facebook.com/aisangam4you...
Watch video || Testing custom Object detector with tensorflow object detection api || online without registration, duration hours minute second in high quality. This video was added by user Al Sangam 30 May 2019, don't forget to share it with your friends and acquaintances, it has been viewed on our site 851 once and liked it 5 people.