Background: The immunohistochemical detection of estrogen (ER) and progesterone (PR) receptors in breast cancer is routinely used for prognostic and predictive testing. Whole slide digitalization supported by dedicated software tools allows quantization of the image objects (e.g. cell membrane, nuclei) and an unbiased analysis of immunostaining results. Validation studies of image analysis applications for the detection of ER and PR in breast cancer specimens provided strong concordance between the pathologist's manual assessment of slides and scoring performed using different software applications. Methods: The effectiveness of two connected semi-automated image analysis software (NuclearQuant v. 1.13 application for Pannoramic™ Viewer v. 1.14) for determination of ER and PR status in formalin-fixed paraffin embedded breast cancer specimens immunostained with the automated Leica Bond Max system was studied. First the detection algorithm was calibrated to the scores provided an independent assessors (pathologist), using selected areas from 38 small digital slides (created from 16 cases) containing a mean number of 195 cells. Each cell was manually marked and scored according to the Allred-system combining frequency and intensity scores. The performance of the calibrated algorithm was tested on 16 cases (14 invasive ductal carcinoma, 2 invasive lobular carcinoma) against the pathologist's manual scoring of digital slides. Results: The detection was calibrated to 87 percent object detection agreement and almost perfect Total Score agreement (Cohen's kappa 0.859, quadratic weighted kappa 0.986) from slight or moderate agreement at the start of the study, using the un-calibrated algorithm. The performance of the application was tested against the pathologist's manual scoring of digital slides on 53 regions of interest of 16 ER and PR slides covering all positivity ranges, and the quadratic weighted kappa provided almost perfect agreement (κ = 0.981) among the two scoring schemes. Conclusions: NuclearQuant v. 1.13 application for Pannoramic™ Viewer v. 1.14 software application proved to be a reliable image analysis tool for pathologists testing ER and PR status in breast cancer.

Quality inspection is inevitable in the steel industry so there are already benchmark datasets for the visual inspection of steel surface defects. In our work, we show, contrary to previous recent articles, that a generic state-of-art deep neural network is capable of almost-perfect classification of defects of two popular benchmark datasets. However, in real-life applications new types of errors can always appear, thus incremental learning, based on very few example shots, is challenging. In our article, we address the problems of the low number of available shots of new classes, the catastrophic forgetting of known information when tuning for new artifacts, and the long training time required for re-training or fine-tuning existing models. In the proposed new architecture we combine EfficientNet deep neural networks with randomized classifiers to aim for an efficient solution for these demanding problems. The classification outperforms all other known approaches, with an accuracy 100% or almost 100%, on the two datasets with the off-the-shelf network. The proposed few-shot learning approach shows considerably higher accuracy at a low number of shots than the different methods under testing, while its speed is significantly (at least 10 times) higher than its competitors. According to these results, the classification and few-shot learning of steel surface defects can be solved more efficiently than was possible before.

Pill image recognition by machine vision can reduce the risk of taking the wrong medications, a severe healthcare problem. Automated dispensing machines or home applications both need reliable image processing techniques to compete with the problem of changing viewing conditions, large number of classes, and the similarity in pill appearance. The problem is attacked with a multi‐stream, two‐phase metric embedding neural model. To enhance the metric learning procedure, dynamic margin setting is introduced into the loss function. Moreover, it is shown that besides the visual features of drug samples, even free text of drug leaflets (processed with a natural language model) can be used to set the value of the margin in the triplet loss and thus increase the recognition accuracy of testing. Thus, besides using the conventional metric learning approach, the given discriminating features can be explicitly injected into the metric model using the NLP of the free text of pill leaflets or descriptors of images of selected pills. The performance on two datasets is analysed and a 1.6% (two‐sided) and 2.89% (one‐sided) increase in Top‐1 accuracy on the CURE dataset is reported compared to existing best results. The inference time on CPU and GPU makes the proposed model suitable for different kinds of applications in medical pill verification; moreover, the approach applies to other areas of object recognition where few‐shot problems arise. The proposed high‐level feature injection method (into a low‐level metric learning model) can also be exploited in other cases, where class features can be well described with textual or visual cues.

The analysis of motion information is one of the main tools for the understanding of complex behaviors in video. However, due to the quality of the optical flow of low-cost surveillance camera systems and the complexity of motion, new robust image-processing methods are required to generate reliable higher-level information. In our novel approach there is no need for tracking objects (vehicles, pedestrians) in order to recognize anomalous motion, but dense optical flow information is used to construct mixtures of Gaussians, which are analyzed temporally. We create a multilevel model, where low-level states of non-overlapping image regions are modeled by continuous hidden Markov models (HMMs). From low-level HMMs we compose high-level HMMs to analyze the occurrence of the low-level states. The processing of large numbers of data in traditional HMMs can result in a precision problem due to the multiplication of low probability values. Thus, besides introducing new motion models, we incorporate a scaling technique into the mathematical model of HMMs to avoid precision problems and to get an effective tool for the analysis of large numbers of motion vectors. We illustrate the use of our models with real-life traffic videos.

In this paper, we present an analysis of low complexity signal processing algorithms capable of identifying special noises, such as the sounds of forest machinery (used for forestry, logging). Our objective is to find methods that are able to detect internal combustion engines in rural environment, and are also easy to implement on low power devices of WSNs (wireless sensor networks). In this context, we review different methods for detecting illegal logging, with an emphasis on autocorrelation and TESPAR audio techniques. The processing of extracted audio features is to be solved with limited memory and processor resources typical for low cost sensors modes. The representation of noise models is also considered with different archetypes. Implementations of the proposed methods were tested not by simulations but on sensor nodes equipped with an omnidirectional microphone and a low power microcontroller. Our results show that high recognition rate can be achieved using time domain algorithms and highly energy efficient and inexpensive architectures.

Separating the moving image parts from the static background is an important phase in video surveillance applications. The method based on mixture of Gaussians (MOG) is an often used and robust approach to learn the background automatically and adaptively. Known MOG methods often suffer from the phenomena called the foreground aperture problem, when parts of large moving homogenous regions become part of the background instead of being selected as moving pixels. This article introduces a new method to eliminate this problem.