Numerous changes in algorithms have been observed by object detection to enhance both speed and accuracy. In this research, we present a method to improve the behavioral clone of self-driving cars. Thus, we first crea...
详细信息
Numerous changes in algorithms have been observed by object detection to enhance both speed and accuracy. In this research, we present a method to improve the behavioral clone of self-driving cars. Thus, we first create a collection of videos and information required for safe driving on different routes and conditions. The detection of obstacles is done with the proposed algorithm called "YOLO non-maximumsuppression fuzzy algorithm, which performs the driver reaction to obstacles with greater accuracy and more speed than the obstacles detection algorithms using the designed framework. The network is trained by the driver's performance, and hence, the output used to control the vehicle is obtained. The non-maximum suppression algorithm plays an essential role in object detection and tracking. An effective hybrid method of fuzzy and NMS algorithms is provided in this paper to improve the problem mentioned. The proposed method improves the average accuracy of the detection network. The performance of the designed algorithm was examined using two different types of KITTI data and the data collected using the personal vehicle and the data we gathered. The proposed algorithm was assessed with evaluation accuracy criteria, which revealed that the method has a higher speed (above 64.41%), a lower FPR (below 6.89%), and a lower FNR (below 3.95%) compared with the baseline YOLOv3 model. According to the loss function, the accuracy rate of the network performance is 95%, implying that we have achieved good results.
Computer-aided detection based on Machine Learning (ML) techniques is increasingly used to detect early-stage colorectal polyps from colonoscopy images. This study presents an efficient ML algorithm that analyses colo...
详细信息
Computer-aided detection based on Machine Learning (ML) techniques is increasingly used to detect early-stage colorectal polyps from colonoscopy images. This study presents an efficient ML algorithm that analyses colonoscopy images and accurately detects polyps for reliable diagnosis of the early stages of Colo-Rectal Cancer (CRC). The proposed approach consists of mainly image enhancement, which enhances the low illumination colonoscopy images, followed by feature extraction using Discrete Orthonormal Stockwell Transform (DOST) and classification by a Support Vector Machine (SVM) classifier. We present an efficient image enhancement algorithm that highlights the clinically significant features in the colonoscopy image and the DOST feature extraction method to discriminate between the polyp area and non-polyp region in the colonoscopy data. The proposed method has been trained using the publicly available databases CVC ClinicDB and tested using ETIS Larib and CVC ColonDB. A sliding window with NMS-based post-processing is used in the selection of polyps from the test images. The performance measures are found in terms of precision (93.76%), recall (92.71%), F1 score (93.23%) and F2 score (93.54%) for CVC ColonDB database and precision (80.97%), recall (93.12%), F1 score (86.62%) and F2 score (83.13%) for the ETIS Larib database. Comparison with the existing method shows that the proposed approach surpasses the existing one in terms of precision, recall, F1-score, F2-score in the CVC ColonDB, and in terms of recall, F1-score in the Etis Larib database. This method would help doctors with timely evaluation and analysis of anomalies from colonoscopy data, which would help in the early planning of preventive or therapeutic protocols.
The non-maximumsuppression (NMS) algorithm, which merges neighboring bounding boxes as the detection result, is widely utilized in object detection methods. However, the traditional NMS algorithm is not suitable for ...
详细信息
ISBN:
(纸本)9781665491822
The non-maximumsuppression (NMS) algorithm, which merges neighboring bounding boxes as the detection result, is widely utilized in object detection methods. However, the traditional NMS algorithm is not suitable for text detection in natural scenes, especially for long texts and dense texts. In this paper, we observe that the coordinates of candidate bounding boxes are presented in skew distribution. On this observation, an improved NMS algorithm called SD-NMS (Skew Distribution NMS) is designed. First, the mode and the median of the coordinate set are counted to filter out the redundant bounding boxes. Then the left bounding boxes are merged for the location of the text regions. The SD-NMS method improves the detection ability of the model without extra model training and can be easily embedded in text detection methods. The experimental results show that our method obtains F-measure more than other NMS methods in public data sets ICDAR2015 and MSRATD500.
Face detection is one of the most popular topics in computer vision. There are several well-known techniques for face detection, such as the Viola-Jones detector. However, the performance of the Viola-Jones detector i...
详细信息
ISBN:
(纸本)9781538644584
Face detection is one of the most popular topics in computer vision. There are several well-known techniques for face detection, such as the Viola-Jones detector. However, the performance of the Viola-Jones detector is limited since it mainly applies the simple Haar-based features. Many advanced methods, especially the convolutional neural network (CNN) based method, have very good performance in face detection. However, they require huge amount of training data. Moreover, most of existing algorithms are not robust to rotation, head-up, and head-down cases. In this paper, we find that, with some modifications, the Viola-Jones detector can also have very good performance in face detection. In addition to the Haar features, we also apply the prominent features and the color information. With the contour information, the edge-aware filter, the background smoother, the fuzzy classifier, and the relative locations, the prominent features, such as eyes, mouths, noses, and ears, can be extracted accurately. With these features, the accuracy of face detection can be much improved. Simulations show that, even if huge amount of training data is not applied, the proposed algorithm has better performance than state-of-the-art face detection methods, including the CNN-based method.
暂无评论