Posts

Object Detection: Previous Works

-
Image
Here in this post, we will discuss previous models of object detection, conventional and state of the art object detection models. VGG-16 VGG16 also known as  OxfordNet , is a convolutional neural network(CNN) model. This model was presented by Simonyan and Zisserman in their paper "Very Deep Convolutional Networks for Large Scale Image Recognition". The 16 in the VGG16 stands for the 16 layers, which contain weights. This model can achieve 92.7% top-5 test accuracy in  ImageNet , that is a dataset of more than 14 million images which approximately contains more than 20,000 categories and belongs to 1000 classes. The dataset, ImageNet comprises of images with varying-resolution, hence the images need to be down-sampled to a specific-resolution of 256x256. Here's the architecture of VGG16 model. A distinctive quality of VGG16 is that it doesn't have a large number of hyper-parameters, hence making a refinement over AlexNet. Instead it has convolution layers of 3x3 filt...
Post a Comment
Read more

Object Detection: Basics of Computer Vision

-
Image
Object Detection, localizing objects in image frames, has been the focus of the Deep Learning, and Computer vision fraternity for decades now.  It plays a crucial role in the autonomous industry and the algorithms here can be used for Image classification, which indeed can be used almost everywhere. Basics Starting with the basics of Computer Vision,  first, we need to know about convolution filters and Image transformations. 1. Vertical Edge Detection Filter: As the name suggests, detecting vertical edges in an image. This can be an important filter when you want to detect something which has clear vertical lines, like boxes, sudoku grid, tables, and graphs. Example:  [[1,0,-1], [1,0,-1], [1,0,-1]],  Sobel Filter and Scharr Filter. 2. Padding: Padding is basically adding pixels to the borders. It is a really helpful tool when it comes to changing the image size to fit the input size of the layers/nodes. It helps in avoiding the Shrinking Problem and information...
Post a Comment
Read more

Object Detection in Twilight: Introduction

-
PES University – Intel India Technovation Contest 2020 CIE in collaboration with Department of CS & ECE  The purpose of the activity would be to enhance the vision of the driver with other sensors or smarter algorithms that can work in low light scenarios – dusk, night, in lights from headlights, etc. The solution would be able to detect common objects like cars, two-wheelers, auto-rickshaws, pedestrians, riders in the night/low-light conditions. The system can be developed to work in the visible spectrum, infra-red, or ultra-sound. Target performance for an ideal scenario would have the algorithm running at 30Hz. This project was carried out by a team of students from PES University- Kashish Oberoi Shobhit Tuli Anuvrat Singhal Anushk Misra  Under the Guidance of Prof. Veena S., Department of ECE, PES University.
Post a Comment
Read more