Real Time Object Detection using CNN

Akash Tripathi,Tarun Kanth Dhansetty,J Selva Kumar,T V. Ajay Kumar

doi:10.14419/ijet.v7i2.24.11994

Akash Tripathi, Tarun Kanth Dhansetty + Show 2 more

Open Access

https://doi.org/10.14419/ijet.v7i2.24.11994

Copy DOI

Abstract

Achieving new heights in object detection and image classification was made possible because of Convolution Neural Network(CNN). However, compared to image classification the object detection tasks are more difficult to analyze, more energy consuming and computation intensive. To overcome these challenges, a novel approach is developed for real time object detection applications to improve the accuracy and energy efficiency of the detection process. This is achieved by integrating the Convolutional Neural Networks (CNN) with the Scale Invariant Feature Transform (SIFT) algorithm. Here, we obtain high accuracy output with small sample data to train the model by integrating the CNN and SIFT features. The proposed detection model is a cluster of multiple deep convolutional neural networks and hybrid CNN-SIFT algorithm. The reason to use the SIFT featureis to amplify the model‟s capacity to detect small data or features as the SIFT requires small datasets to detect objects. Our simulation results show better performance in accuracy when compared with the conventional CNN method. As the resources like RAM, graphic card, ROM, etc. are limited we propose a pipelined implementation on an aggregate Central Processing Unit(CPU) and Graphical Processing Unit(GPU) platform.

Highlights

Achieving new heights in object detection and image classification was made possible because of Convolution Neural Network(CNN)
Some of the techniques algorithms have been developed for object detection tasks. include algorithms like SVM‟s(Support vector machines), Numerous handcrafted feature extraction techniques for object stochastic gradient descent (SGD)(Stochastic Gradient Descent), Convolution Neural Support detection such as SVM(Support Vector Machines), Vector Machines (CNSVMs) which can be found in SGD(Stochastic Gradient Descent), Convolutional Neural Support [10],[11],[12],[13].In [11], the author suggests that support vector
We propose a hybrid Convolutional Neural Network and Scale Invariant Feature Transform (SIFT) aggregator for efficient object detection

Summary

Design and Architecture

We are using a webcam of eight mega pixels, for the system to be more robust against the noise. These transformations include rotated and scaled trained images In this model we have built a characteristic CNN network architecture from square one. The SIFT [7] algorithm is used to extract the key points from the object image. After detecting features like key points, magnitude and direction of the object image, using key points of neighboring pixels we calculate the image gradient. Another example of SIFT and CNN is proposed in [8]. There‟s a use the Euclidean distance to match them regardless of the key fully connected layer where the highest level of reasoning in the points' scale. Each new image feed for test is compared individually to the database and

Result

35 Table 1

Conclusion