We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Automatic Image Cropping for Online Classifieds | Alexey Grigorev
Automatic image cropping for online classifieds using neural networks and silence detection, with promising results in various categories and scalability considerations.
- Automatic image cropping for online classifieds is essential for enhancing user experience and increasing engagement.
- The speaker’s approach uses a combination of neural networks and silency detection to crop images.
- The model is tested on various categories, including cars, animals, and fashion, and shows promising results.
- The speaker discusses the challenges of scaling the model to process large amounts of data and the need for a robust infrastructure.
- Kubernetes is used to scale the model and handle high traffic.
- The speaker mentions that the model is still a work in progress and that further testing and refinement are needed.
- The use of convolutional neural networks (CNNs) is discussed, and the speaker highlights their ability to process images efficiently.
- The speaker also mentions the concept of Deeply Supervised Skip Connections Network (DSS) and its application in image cropping.
- The talk concludes with the speaker emphasizing the importance of collaboration between different business entities to develop and implement innovative solutions.