[Application Test of the AI-Automatic Diagnostic System for Ki-67 in Breast Cancer].

OBJECTIVE: To study the different methods of artificial intelligence (AI)-assisted Ki-67 scoring of clinical invasive ductal carcinoma (IDC) of the breast and to compare the results. METHODS: A total of 100 diagnosed IDC cases were collected, including slides of HE staining and immunohistochemical Ki-67 staining and diagnosis results. The slides were scanned and turned into whole slide image (WSI), which were then scored with AI. There were two AI scoring methods. One was fully automatic counting by AI, which used the scoring system of Ki-67 automatic diagnosis to do counting with the whole image of WSI. The second method was semi-automatic AI counting, which required manual selection of areas for counting, and then relied on an intelligent microscope to conduct automatic counting. The diagnostic results of pathologists were taken as the results of pure manual counting. Then the Ki-67 scores obtained by manual counting, semi-automatic AI counting and automatic AI counting were pairwise compared. The Ki-67 scores obtained from the manual counting (pathological diagnosis results), semi-automatic AI and automatic AI counts were pair-wise compared and classified according to three levels of difference: difference ≤10%, difference of >10%-<30% and difference ≥30%. Intra-class correlation coefficient ( ICC) was used to evaluate the correlation. RESULTS: The automatic AI counting of Ki-67 takes 5-8 minutes per case, the semi-automatic AI counting takes 2-3 minutes per case, and the manual counting takes 1-3 minutes per case. When results of the two AI counting methods were compared, the difference in Ki-67 scores was all within 10% (100% of the total), and the ICC index being 0.992. The difference between manual counting and semi-automatic AI was less than 10% in 60 cases (60% of the total), between 10% and 30% in 37 cases (37% of the total), and more than 30% in only 3 cases (3% of the total), ICC index being 0.724. When comparing automatic AI with manual counting, 78 cases (78% of the total) had a difference of ≤10%, 17 cases (17% of the total) had a difference of between 10% and 30%, and 5 cases (5%) had a difference of ≥30%, the ICC index being 0.720. The ICC values showed that there was little difference between the results of the two AI counting methods, indicating good repeatability, but the repeatability between AI counting and manual counting was not particularly ideal. CONCLUSION: AI automatic counting has the advantage of requiring less manpower, for the pathologist is involved only for the verification of the diagnosis results at the end. However, the semi-automatic method is better suited to the diagnostic habits of pathologists and has a shorter turn-over time compared with that of the fully automatic AI counting method. Furthermore, in spite of its higher repeatability, AI counting, cannot serve as a full substitute for pathologists, but should instead be viewed as a powerful auxiliary tool.

Automatic Outcome in Manual Dexterity Assessment Using Colour Segmentation and Nearest Neighbour Classifier.

Objective assessment of motor function is an important component to evaluating the effectiveness of a rehabilitation process. Such assessments are carried out by clinicians using traditional tests and scales. The Box and Blocks Test (BBT) is one such scale, focusing on manual dexterity evaluation. The score is the maximum number of cubes that a person is able to displace during a time window. In a previous paper, an automated version of the Box and Blocks Test using a Microsoft Kinect sensor was presented, and referred to as the Automated Box and Blocks Test (ABBT). In this paper, the feasibility of ABBT as an automated tool for manual dexterity assessment is discussed. An algorithm, based on image segmentation in CIELab colour space and the Nearest Neighbour (NN) rule, was developed to improve the reliability of automatic cube counting. A pilot study was conducted to assess the hand motor function in people with Parkinson's disease (PD). Three functional assessments were carried out. The success rate in automatic cube counting was studied by comparing the manual (BBT) and the automatic (ABBT) methods. The additional information provided by the ABBT was analysed to discuss its clinical significance. The results show a high correlation between manual (BBT) and automatic (ABBT) scoring. The lowest average success rate in cube counting for ABBT was 92%. Additionally, the ABBT acquires extra information from the cubes' displacement, such as the average velocity and the time instants in which the cube was detected. The analysis of this information can be related to indicators of health status (coordination and dexterity). The results showed that the ABBT is a useful tool for automating the assessment of unilateral gross manual dexterity, and provides additional information about the user's performance.

Potential application of digital image-processing method and fitted logistic model to the control of oriental fruit moths (Grapholita molesta Busck).

Monitoring of oriental fruit moths (Grapholita molesta Busck) is a prerequisite for its control. This study introduced a digital image-processing method and logistic model for the control of oriental fruit moths. First, five triangular sex pheromone traps were installed separately within each area of 667 m2 in a peach orchard to monitor oriental fruit moths consecutively for 3 years. Next, full view images of oriental fruit moths were collected via a digital camera and then subjected to graying, separation and morphological analysis for automatic counting using MATLAB software. Afterwards, the results of automatic counting were used for fitting a logistic model to forecast the control threshold and key control period. There was a high consistency between automatic counting and manual counting (0.99, P < 0.05). According to the logistic model, oriental fruit moths had four occurrence peaks during a year, with a time-lag of 15-18 days between adult occurrence peak and the larval damage peak. Additionally, the key control period was from 28 June to 3 July each year, when the wormy fruit rate reached up to 5% and the trapping volume was approximately 10.2 per day per trap. Additionally, the key control period for the overwintering generation was 25 April. This study provides an automatic counting method and fitted logistic model with a great potential for application to the control of oriental fruit moths.

Large-scale performance assessment of the BG-Counter 2 used with two different mosquito traps.

BACKGROUND: Mosquitoes are important vectors of pathogens. They are usually collected with CO2-baited traps and subsequently identified by morphology. This procedure is very time-consuming. Automatic counting traps could facilitate timely evaluation of the local risk for mosquito-borne pathogen transmission or decision-making on vector control measures, but the counting accuracy of such devices has rarely been validated in the field. METHODS: The Biogents (BG)-Counter 2 automatically counts mosquitoes by discriminating the size of captured objects directly in the field and transmits the data to a cloud server. To assess the accuracy of this counting device, 27 traps were placed at 19 sampling sites across Germany and used in daily, weekly or bimonthly intervals from April until October 2021. The BG-Counter 2 was attached to a CO2-trap (BG-Pro trap = CO2-Pro) and the same trap was converted to also attract gravid mosquitoes (upside-down BG-Pro trap with a water container beneath = CO2-Pro-gravid). All captured mosquitoes were identified by morphology. The number of females (unfed and gravid), mosquito diversity and the number of identified specimens in relation to the counting data of the BG-Counter were compared between the two trapping devices to evaluate sampling success and counting accuracy. RESULTS: In total 26,714 mosquitoes were collected during 854 trap days. The CO2-Pro-gravid trap captured significantly more mosquitoes per trap day for all specimens, gravid females and non-gravid females, while there was no difference in the mosquito diversity. The linear model with the captured mosquitoes as a response and the counted specimens as a predictor explained only a small degree of the variation within the data (R2 = 0.16), but per individual trap the value could reach up to 0.62 (mean R2 = 0.23). The counting accuracy for the daily samples had a significant positive correlation with sample size, resulting in higher accuracy for the CO2-Pro-gravid trap and higher accuracy for sites and sampling months with high mosquito abundance. CONCLUSIONS: While the accuracy of the BG-Counter 2 is quite low, the device is able to depict mosquito phenology and provide information about local population dynamics.

Automatic Detection of Feral Pigeons in Urban Environments Using Deep Learning.

The overpopulation of feral pigeons in Hong Kong has significantly disrupted the urban ecosystem, highlighting the urgent need for effective strategies to control their population. In general, control measures should be implemented and re-evaluated periodically following accurate estimations of the feral pigeon population in the concerned regions, which, however, is very difficult in urban environments due to the concealment and mobility of pigeons within complex building structures. With the advances in deep learning, computer vision can be a promising tool for pigeon monitoring and population estimation but has not been well investigated so far. Therefore, we propose an improved deep learning model (Swin-Mask R-CNN with SAHI) for feral pigeon detection. Our model consists of three parts. Firstly, the Swin Transformer network (STN) extracts deep feature information. Secondly, the Feature Pyramid Network (FPN) fuses multi-scale features to learn at different scales. Lastly, the model's three head branches are responsible for classification, best bounding box prediction, and segmentation. During the prediction phase, we utilize a Slicing-Aided Hyper Inference (SAHI) tool to focus on the feature information of small feral pigeon targets. Experiments were conducted on a feral pigeon dataset to evaluate model performance. The results reveal that our model achieves excellent recognition performance for feral pigeons.

Automatic Detection of Moths (Lepidoptera) with a Funnel Trap Prototype.

Monitoring insect populations is essential to optimise pest control with the correct protection timing and the avoidance of unnecessary insecticide use. Modern real-time monitoring practices use automatic insect traps, which are expected to be able to estimate the population sizes of pest animals with high species specificity. There are many solutions to overcome this challenge; however, there are only a few data that consider their accuracy under field conditions. This study presents an opto-electronic device prototype (ZooLog VARL) developed by us. A pilot field study evaluated the precision and accuracy of the data filtering using an artificial neural network(ANN) and the detection accuracy of the new probes. The prototype comprises a funnel trap, sensor-ring, and data communication system. The main modification of the trap was a blow-off device that prevented the escape of flying insects from the funnel. These new prototypes were tested in the field during the summer and autumn of 2018, detecting the daily and monthly flight of six moth species (Agrotis segetum, Autographa gamma, Helicoverpa armigera, Cameraria ohridella, Grapholita funebrana, Grapholita molesta). The accuracy of ANN was always higher than 60%. In the case of species with larger body sizes, it reached 90%. The detection accuracy ranged from 84% to 92% on average. These probes detected the real-time catches of the moth species. Therefore, weekly and daily patterns of moth flight activity periods could be compared and displayed for the different species. This device solved the problem of multiple counting and gained a high detection accuracy in target species cases. ZooLog VARL probes provide the real-time, time-series data sets of each monitored pest species. Further evaluation of the catching efficiency of the probes is needed. However, the prototype allows us to follow and model pest dynamics and may make more precise forecasts of population outbreaks.

Using DeepLab v3 + -based semantic segmentation to evaluate platelet activation.

This research used DeepLab v3 + -based semantic segmentation to automatically evaluate the platelet activation process and count the number of platelets from scanning electron microscopy (SEM) images. Current activated platelet recognition and counting methods include (a) using optical microscopy or SEM images to identify and manually count platelets at different stages, or (b) using flow cytometry to automatically recognize and count platelets. However, the former is time- and labor-consuming, while the latter cannot be employed due to the complicated morphology of platelet transformation during activation. Additionally, because of how complicated the transformation of platelets is, current blood-cell image analysis methods, such as logistic regression or convolution neural networks, cannot precisely recognize transformed platelets. Therefore, this study used DeepLab v3 + , a powerful learning model for semantic segmentation of image analysis, to automatically recognize and count platelets at different activation stages from SEM images. Deformable convolution, a pretrained model, and deep supervision were added to obtain additional platelet transformation features and higher accuracy. The number of activated platelets was predicted by dividing the segmentation predicted platelet area by the average platelet area. The results showed that the model counted the activated platelets at different stages from the SEM images, achieving an error rate within 20%. The error rate was approximately 10% for stages 2 and 4. The proposed approach can thus save labor and time for evaluating platelet activation and facilitate related research.

Development and validation of software that quantifies the larval mortality of Rhipicephalus (Boophilus) microplus cattle tick.

Rhipicephalus (Boophilus) microplus is controlled almost exclusively using synthetic acaricides, and reports of resistant populations have been described worldwide. Several time-consuming and laborious toxicological in vitro tests have been used to diagnose acaricidal resistance, especially those that require differential counting of live and dead larvae. Larval mortality is currently done manually and subjectively, which can limit the performance of a large number of tests and comparing results between different laboratories. The present study aimed to develop and validate a new automatic counting method to evaluate tick larval mortality. A software for differentiation of live and dead larvae was developed using different steps: obtaining videos; image segmentation using the firefly algorithm; detection of larvae with the fast radial symmetry transform technique (FRST); and tracking of the larvae at a given time. Larval immersion tests with ivermectin, cypermethrin, and fipronil were performed to validate the developed software. The larval mortality evaluation was performed by (1) recording for 60 s for each package and (2) manual counts of the same sample using three different analysts, each responsible for counting one replicate of each test. All videos obtained were copied and cut at 60, 40, and 20 s for later analysis in the counting software. The median lethal doses (LD50) of the different compounds in each test were calculated for each method (automatic and manual) for different video times. There was no statistical difference in LD50 between manual and automatic count techniques for ivermectin and fipronil. The LD50 of cypermethrin calculated with manual evaluation was up to 2.2 times lower than that of automatic evaluation. The acquisition time of the videos was 2.9-4.4 times faster than the manual evaluation. The average processing time for each video was 5.73 min, regardless of their duration. Thus, the method developed for automatic counting of tick larvae was validated, and although it still has points to be optimized, it can be considered a viable alternative for determining the percentage of tick larvae mortality and could be applied to toxicological in vitro tests with acaricides, assisting in the diagnosis of resistant tick populations and studies of novel acaricide development.

An Integrated Goat Head Detection and Automatic Counting Method Based on Deep Learning.

Goat farming is one of the pillar industries for sustainable development of national economies in some countries and plays an active role in social and economic development. In order to realize the precision and intelligence of goat breeding, this paper describes an integrated goat detection and counting method based on deep learning. First, we constructed a new dataset of video images of goats for the object tracking task. Then, we took YOLOv5 as the baseline of the object detector and improved it using a series of advanced methods, including: using RandAugment to explore suitable data augmentation strategies in a real goat barn environment, using AF-FPN to improve the network's ability to represent multi-scale objects, and using the Dynamic Head framework to unify the attention mechanism with the detector's heads to improve its performance. The improved detector achieved 92.19% mAP, a significant improvement compared to the 84.26% mAP of the original YOLOv5. In addition, we also input the information obtained by the detector into DeepSORT for goat tracking and counting. The average overlap rate of our proposed method is 89.69%, which is significantly higher than the 82.78% of the original combination of YOLOv5 and DeepSORT. In order to avoid double counting as much as possible, goats were counted using the single-line counting based on the results of goat head tracking, which can support practical applications.

Segmentation of nearly isotropic overlapped tracks in photomicrographs using successive erosions as watershed markers.

INTRODUCTION: Procedures for measuring and counting tracks are time-consuming and involve practical problems. The precision of automatic counting methods is not satisfactory yet; the major challenges are distinguishing tracks and material defects, identifying small tracks and defects of similar size, and detecting overlapping tracks. MATERIALS AND METHODS: Here, we address the overlapping tracks issue using the algorithm Watershed Using Successive Erosions as Markers (WUSEM), which combines the watershed transform, morphological erosions and labeling to separate regions in photomicrographs. We tested this method in two data sets of diallyl phthalate (DAP) photomicrographs and compared the results when counting manually and using the classic watershed and H-watershed transforms. RESULTS: The mean automatic/manual efficiency counting ratio when using WUSEM in the test data sets is 0.97 ± 0.11. CONCLUSION: WUSEM shows reliable results when used in photomicrographs presenting almost isotropic objects. Also, diameter and eccentricity criteria may be used to increase the reliability of this method.

Development and preparation of active starch films carrying tea polyphenol.

Starch films incorporated with tea polyphenol (TP) were developed to produce active food packaging. The effect of the incorporation of TP with different content on the structure, physicochemical properties, antioxidant activity and antimicrobial activity of the starch films was systematically evaluated. Results showed that TP was well dispersed in the starch matrix, which induced a slight influence on the surface and barrier properties of the films. TP addition led to an important improvement in antioxidant capability, as well as inhibition efficiency against the microorganisms of S. aureus and E. coli. However, a decrease in mechanical properties of films was observed. Moreover, a new automatic counting method which combined the computer vision and machine learning algorithm was developed to identify and count the colonies, and the method performed much faster without subjective uncertainty.

Automatic counting of microglial cell activation and its applications.

Glaucoma is a multifactorial optic neuropathy characterized by the damage and death of the retinal ganglion cells. This disease results in vision loss and blindness. Any vision loss resulting from the disease cannot be restored and nowadays there is no available cure for glaucoma; however an early detection and treatment, could offer neuronal protection and avoid later serious damages to the visual function. A full understanding of the etiology of the disease will still require the contribution of many scientific efforts. Glial activation has been observed in glaucoma, being microglial proliferation a hallmark in this neurodegenerative disease. A typical project studying these cellular changes involved in glaucoma often needs thousands of images - from several animals - covering different layers and regions of the retina. The gold standard to evaluate them is the manual count. This method requires a large amount of time from specialized personnel. It is a tedious process and prone to human error. We present here a new method to count microglial cells by using a computer algorithm. It counts in one hour the same number of images that a researcher counts in four weeks, with no loss of reliability.

Sistema de aquisição e processamento de imagens de ovitrampas para o combate a dengue / System for acquisition and processing of ovitraps images to fight dengue

A dengue é uma doença viral transmitida ao homem pela picada do mosquito Aedes aegypti contaminado. A erradicação do mosquito é extremamente difícil. Assim a implementação e o acompanhamento das medidas de controle da população do inseto são essenciais. O estabelecimento de métodos de monitoração do vetor da doença é uma preocupação de vários países tropicais. No estado de Pernambuco (Brasil), a ocorrência do vetor vem sendo monitorada pelo uso de ovitrampas, armadilhas especiais para a deposição e contagem dos ovos do mosquito. No entanto, comumente esta contagem é realizada manualmente com o auxílio de lupa ou microscópio, método trabalhoso que exige tempo de profissionais treinados e está sujeita a variações. Este trabalho apresenta uma ferramenta capaz de adquirir e armazenar imagens das palhetas das ovitrampas, realizar a contagem semi-automática e automática dos ovos, sem a utilização do microscópio. O sistema desenvolvido é baseado em uma plataforma óptica, uma interface homem-máquina e um software de aquisição de imagem, com a contagem assistida dos ovos do mosquito. Esta contagem semi-automática gerou um ganho de velocidade na contagem de três vezes. As informações obtidas pelo sistema são enviadas pela Internet para um computador servidor WEB, onde são analisadas por técnicas de processamento de imagens. A contagem automática dos ovos baseia-se nos processos de segmentação, filtragem e quantificação. Este método foi aplicado em um conjunto de 100 imagens obtendo um erro global de 2,67%. Dois protótipos do sistema foram instalados e implementados, em duas diferentes cidades do estado de Pernambuco.

Dengue is a viral disease transmitted to humans by the bite of the infected Aedes aegypti mosquito. The Dengue vector eradication is extremely difficult, therefore the implementation and evaluation of public policies are important issues. New methods of monitoring disease vectors are major concern in many tropical countries. In the state of Pernambuco (Brazil), the presence of the vector has been monitored by the use of ovitraps, special traps for the mosquito eggs deposition, and eggs counting methods. One drawback of the monitoring procedure is that the egg counting method has been done manually with a magnifying glass or microscope, a laborious method requiring time of trained personnel and is subject to variations. This work presents a tool capable of acquiring and storing images of the ovitraps palettes and counting eggs, semi-automatically and automatically. The developed system is based on an optical platform, a man-machine interface, and a software for mosquito eggs counting. This semi-automatic count generated a three-fold increase in the counting speed. The obtained information is sent over the Internet to a WEB server computer, where it is analyzed using image processing techniques. The automatic counting procedure is based on segmentation, filtering, and quantification processes. This method was applied to a group of 100 images giving a total error of 2.67%. Two prototypes of the device have been installed and implemented in two different cities in Pernambuco state.