Objects on images indexing

Author

Anastasiia Dolgaryeva

Delivery Manager

    There is an enormous amount of visual information in the world. The most actual task is to find exactly what you need. Instead of Google image search, there are a lot of specific tasks with a limited number of object types for image tagging and indexing.

    Let’s look at the computer vision technologies usage example for detecting wedding accessories and classifying them, providing information on different accessories characteristics provided in the table below.

    wedding accessories classification

    Technology

    The first step of this project is data labeling. The customer has provided a dataset of wedding photos without any labeling. And we used Amazon MTurk service to provide such.

    Amazon pipeline consists оf :

    1. HTML template, which workers used for drawing bounding boxes and selecting the label.
    wedding rings detection
    Rings

    The second step was to train a neural network for category detection. We tested several neural net architectures, such as MobileNet, NasNet, ResNet, and YOLO, to find the best accuracy/speed ratio. Finally, the YOLOv2 neural net was trained on the labeled dataset.

    wedding Bouquet detection
    Bouquet                                                              

    The third step was to train an ensemble of models to make a classification of detected categories by tags. We used PyTorch pre-trained vgg16 models to train it on a labeled dataset.

    Bride detection
    Bride

     

    detection characteristics

     

    The demonstrated approach may be used in eCommerce to automate image tagging. It’s also applicable for quick searches of goods with similar characteristics.

    While this project focuses on deep learning for complex object detection, other computer vision challenges can be addressed with different toolsets. For instance, a common task like extracting specific text from a document can be accomplished by combining classic image processing with Optical Character Recognition (OCR).

    Contact us if you have questions or ideas, and follow our blog updates.

    Read more: Multiple object tracking using person re-identification

    Share this post:

    • #AI & Machine Learning
    • #Computer Vision
    • #Data Science
    • #Image recognition
    • #Object detection

    AI & Machine Learning

    Automate repetitive visual monitoring to boost operational efficiency, reduce costly human error, and accelerate core business workflows.

    Check the service
    Related Articles

    Connect with our experts

    Get in touch