In this webinar, we will explore a deep learning library written in Kotlin, learn how to detect objects of different types in images, and create a Kotlin Web Application using Ktor and KotlinDL, which will detect persons, trucks, and cars on the photos.

Speaker: Alexey Zinoviev https://twitter.com/zaleslaw
KotlinDL at GitHub: https://github.com/JetBrains/KotlinDL
Demo: https://github.com/zaleslaw/Ktor-KotlinDL-Object-Detection-examples

Timeline:
0:00 Introduction
3:15 Kotlin for Data Science
4:32 Introduction to KotlinDL
10:23 Image Recognition with KotlinDL
10:41 ModelHubs: TensorFlow Model Hub and ONNX Model Hub
14:04 Difference between ResNet50 and ResNet152
22:29 Object Detection with SSD model
26:50 What is the model’s input shape?
29:08 Object Detection with visualization on Swing JPanel
29:48 Image Preprocessing DSL
32:31 Q&A Part 1
37:01 Client-Server Image Recognition service with Ktor
39:28 Client-Server Object Detection POST request with Ktor
43:17 Single-Page Application for Object Detection with Ktor
46:22 We detected all persons who visited Wall Street one day
52:07 Q&A Part 2

source