Object Class Detection

We describe a web application that takes advantage of new computer vision techniques to allow the user to make searches based on visual similarity of color and texture related to the object of interest. We use a supervised neural network strategy to segment different classes of objects. A strength of this solution is the high speed in generalization of the trained neural networks, in order to obtain an object segmentation in real time. Information about the segmented object, such as color and texture, are extracted and indexed as text descriptions. Our case study is the online commercial offers domain where each offer is composed by text and images. Many successful experiments were done on real datasets in the fashion field.

Object detection is one of the major challenges in visual sensor networks (VSNs) which is set up in the monitoring applications. Many approaches proposed to solve the object detection problem in VSNs, considering diverse metrics such as reliability, energy consumption, detection accuracy and being real-time. In this paper, a survey on the object detection methods in visual sensor networks is presented for the first time. Furthermore, this paper classified the methods precisely. Two main object detection categories in VSNs that explored in this paper are conventional object detection methods and object detection approaches with the camera nodes involvement. To be more precise, presented survey promotes an overview of recent object detection methods' literature with their performance evaluation. Also, this research is challenging and the object detection issue in the visual sensor networks is open caused by differences in estimations and performance metrics. Therefore, the survey concludes with open research challenges.

Computer vision is a process by which we can understand how the images and videos are stored and manipulated, also it helps in the process of retrieving data from either images or videos. Computer Vision is part of Artificial Intelligence. Computer-Vision plays a major role in Autonomous cars, Object detections, robotics, object tracking, etc. OpenCV (Open Source Computer Vision Library) is an open source computer vision and machine learning software library. OpenCV was built to provide a common infrastructure for computer vision applications and to accelerate the use of machine perception in the commercial products. It comes with a highly improved deep learning (dnn) module. This module now supports a number of deep learning frameworks, including Caffe, TensorFlow, and Torch/PyTorch. This does allow us to take our models trained using dedicated deep learning libraries/tools and then efficiently use them directly inside our OpenCV scripts. MediaPipe is a framework mainly used for building audio, video, or any time series data. With the help of the MediaPipe framework, we can build very impressive pipelines for different media processing functions like Multi-hand Tracking, Face Detection, Object Detection and Tracking, etc. I.

Forward Collision Avoidance (FCA) systems in automobiles is an essential part of Advanced Driver Assistance System (ADAS) and autonomous vehicles. These devices currently use, radars as the main sensor. The increasing resolution of camera sensors, processing capability of hardware chipsets and advances in image processing algorithms, have been pushing the camera based features recently. Monocular cameras face the challenge of accurate scale estimation which limits it use as a stand-alone sensor for this application. This paper proposes an efficient system which can perform multi scale object detection which is being patent granted and efficient 3D reconstruction using structure from motion (SFM) framework. While the algorithms need to be accurate it also needs to operate real time in low cost embedded hardware. The focus of the paper is to discuss how the proposed algorithms are designed in such a way that it can be provide real time performance on low cost embedded CPU's which makes use of only Digital Signal processors (DSP) and vector processing cores.

We describe a web application that takes advantage of new computer vision techniques to allow the user to make searches based on visual similarity of color and texture related to the object of interest. We use a supervised neural network strategy to segment different classes of objects. A strength of this solution is the high speed in generalization of the trained neural networks, in order to obtain an object segmentation in real time. Information about the segmented object, such as color and texture, are extracted and indexed as text descriptions. Our case study is the online commercial offers domain where each offer is composed by text and images. Many successful experiments were done on real datasets in the fashion field.

In this paper we propose a semi-supervised multiple instance learning based boosting algorithm for domain adaptation, with face detection as an example. Very often a generic classifier learned using a large volume of training data needs to be tuned to work for a specific scenario. However when deployed, the test scenarios may differ marginally from the training ones. For e.g. a face detection system may be deployed in an airport as well as in an auditorium hallway. The classifier then needs to adapt to the new domain. Instead of retraining the classifier completely using examples from the new scenario, it is desirable to see how much the classifier can "self-learn". Conventional self-learning algorithms consider putative positives on test data given by the base classifier, and select a subset of those based on more stringent thresholds. In this paper we propose an alternative self-learning approach which is based on the popular multiple instance learning approach which makes use of "bags" instead of single instances for training the classifier. We pool the putative positives on a given test image into a positive bag and the putative negatives into a negative bag. We augment this data to the initial training data and retrain the classifier using MILBoost. Specifically the advantage of our approach is that since it makes use of bags it is more robust to classification errors by the base classifier. We demonstrate the improvement in classification accuracy using our approach on Faces in the Wild database. We show that our approach outperforms self-learning and compares favorably with MILBoost trained on manually marked face data without the corresponding increase in labeling effort.