Current state-of-the-art object tracking methods have largely benefited from the public availability of numerous benchmark datasets. However, the focus has been on open-air imagery and much less on underwater visual data. Inherent underwater distortions, such as color loss, poor contrast, and underexposure, caused by attenuation of light, refraction, and scattering, greatly affect the visual quality of underwater data, and as such, existing open-air trackers perform less efficiently on such data. To help bridge this gap, this article proposes a first comprehensive underwater object tracking (UOT100) benchmark dataset to facilitate the development of tracking algorithms well-suited for underwater environments. The proposed dataset consists of 104 underwater video sequences and more than 74 000 annotated frames derived from both natural and artificial underwater videos, with great varieties of distortions. We benchmark the performance of 20 state-of-the-art object tracking algorithms and further introduce a cascaded residual network for underwater image enhancement model to improve tracking accuracy and success rate of trackers. Our experimental results demonstrate the shortcomings of existing tracking algorithms on underwater data and how our generative adversarial network (GAN)-based enhancement model can be used to improve tracking performance. We also evaluate the visual quality of our model's output against existing GAN-based methods using well-accepted quality metrics and demonstrate that our model yields better visual data.

Currently published tone mapping operators (TMO) are often evaluated on a very limited test set of high dynamic range (HDR) images. Thus, the resulting performance index is highly subject to extensive hyperparameter tuning, and many TMOs exhibit sub-optimal performance when tested on a broader spectrum of HDR images. This indicates that there are deficiencies in the generalizable applicability of these techniques. Finally, it is a challenge developing parameter-free tone mapping operators using data-hungry advanced deep learning methods due to the paucity of large scale HDR datasets. In this paper, these issues are addressed through the following contributions: a) a large scale HDR image benchmark dataset (LVZ-HDR dataset) with multiple variations in sceneries and lighting conditions is created to enable performance evaluation of TMOs across a diverse conditions and scenes that will also contribute to facilitate the development of more robust TMOs using state-of-the-art deep learning methods; b) a deep learning-based tone mapping operator (TMO-Net) is presented, which offers an efficient and parameter-free method capable of generalizing effectively across a wider spectrum of HDR content; c) finally, a comparative analysis, and performance benchmarking of 19 state-of-the-art TMOs on the new LVZ-HDR dataset are presented. Standard metrics including the Tone Mapping Quality Index (TMQI), Feature Similarity Index for Tone Mapped images (FSITM), and Natural Image Quality Evaluator (NIQE) are used to qualitatively evaluate the performance index of the benchmarked TMOs. Experimental results demonstrate that the proposed TMO-Net qualitatively and quantitatively outperforms current state-of-the-art TMOs.

The concept of searching and localizing vehicles from live traffic videos based on descriptive textual input has yet to be explored in the scholarly literature. Endowing Intelligent Transportation Systems (ITS) with such a capability could help solve crimes on roadways. One major impediment to the advancement of fine-grain vehicle recognition models is the lack of video testbench datasets with annotated ground truth data. Additionally, to the best of our knowledge, no metrics currently exist for evaluating the robustness and performance efficiency of a vehicle recognition model on live videos and even less so for vehicle search and localization models. In this paper, we address these challenges by proposing V-Localize, a novel artificial intelligence framework for vehicle search and continuous localization captured from live traffic videos based on input textual descriptions. An efficient hashgraph algorithm is introduced to compute valid target information from textual input. This work further introduces two novel datasets to advance AI research in these challenging areas. These datasets include (a) the most diverse and large-scale Vehicle Color Recognition (VCoR) dataset with 15 color classes—twice as many as the number of color classes in the largest existing such dataset—to facilitate finer-grain recognition with color information; and (b) a Vehicle Recognition in Video (VRiV) dataset, a first of its kind video testbench dataset for evaluating the performance of vehicle recognition models in live videos rather than still image data. The VRiV dataset will open new avenues for AI researchers to investigate innovative approaches that were previously intractable due to the lack of annotated traffic vehicle recognition video testbench dataset. Finally, to address the gap in the field, five novel metrics are introduced in this paper for adequately accessing the performance of vehicle recognition models in live videos. Ultimately, the proposed metrics could also prove intuitively effective at quantitative model evaluation in other video recognition applications. T One major advantage of the proposed vehicle search and continuous localization framework is that it could be integrated in ITS software solution to aid law enforcement, especially in critical cases such as of amber alerts or hit-and-run incidents.

Rapid significant weight fluctuations can indicate severe health conditions such as edema due to congestive heart failure or severe dehydration that could require prompt intervention. Daily body weighing does not accurately represent the patient’s body weight fluctuations occurring within a day. The patient’s lack of compliance with tracking their weight measurements is also a predominant issue. Using shoe insole sensors embedded into footwear could achieve accurate real-time monitoring systems for estimating continuous body weight changes. Here, the machine learning models’ predictive capabilities for continuous real-time weight estimation using the insole data are presented. The lack of availability of public datasets to feed these models is also addressed by introducing two novel datasets. The proposed framework is designed to adapt to the patient, considering several unique factors such as shoe type, posture, foot shape, and gait pattern. The proposed framework estimates the mean absolute percentage error of 0.61% and 0.74% and the MAE of 1.009 lbs. and 1.154 lbs. for the less controlled and more controlled experimental settings, respectively. This will help researchers utilize machine learning techniques for more accurate real-time continuous weight estimation using sensor data and enable more reliable aging-in-place monitoring and telehealth.

According to the American cancer society, the average risk of women getting diagnosed with breast cancer during their life is 13%. The World Health Organization also reports that the number of cancer cases is projected to rise to 19.3 million by 2025. Recent research works point out that physicians can only diagnose cancer with 79% accuracy while machine learning procedures achieve 91% accuracy or more. The current challenges are early cancer detection and the efficient and accurate diagnosis of histopathology tissue samples. Several Deep Learning breast cancer classification models have been developed to assist medical practitioners. However, these methods are data hungry and require thousands of training image samples, often coupled with data augmentation to achieve satisfactory results with long training hours. In this paper, we propose a machine learning classification model by integrating the Parameter free Thresholding Adjacency Statistics (PFTAS) with Fibonacci-p patterns for breast cancer detection. Computer simulations on BreakHis cancer datasets in comparison to other machine learning and deep learning-based methods show that (i) the presented method helps eliminate dependence on large training data and data augmentation, (ii) robustness to noise and background stains, and (iii) lightweight model easy to train and deploy.

The growing network of highway video surveillance cameras generates an immense amount of data that proves tedious for manual analysis. Automated real-time analysis of such data may provide many solutions, including traffic monitoring, traffic incident detection, and smart-city planning. More specifically, assessing traffic speed and density is critical in determining dynamic traffic conditions and detecting slowdowns, traffic incidents, and traffic alerts. However, despite several advancements, there are numerous challenges in estimating vehicle speed and traffic density, which are integral parts of ITS. Some of these challenges include variations in road networks, illumination constraints, weather, structure occlusion, and vehicle user-driving behavior. To address these issues, this paper proposes a novel deep learning-based framework for instant-level vehicle speed and traffic flow density estimation to effectively harness the potential of existing large-scale highway surveillance cameras to assist in real-time traffic analysis. This is achieved using the state-of-the-art region-based Siamese MOT network, SiamMOT which detects and associates object instances for multi-object tracking (MOT), to accurately estimate instant level vehicle speed in live video feeds. The UA-DETRAC dataset is used to train the speed estimation model. Computer simulations show that the proposed framework a) allows the classifying of traffic density into light, medium, or heavy traffic flows, b) is robust to different types of road networks and illuminations without prior road information, and c) shows good performance when compared to current state-of-the-art methods using adequate performance metrics.