Development of New Staining Procedures for Diagnosing Cryptosporidium spp. in Fecal Samples by Computerized Image Analysis

Interpretation errors may still represent a limiting factor for diagnosing Cryptosporidium spp. oocysts with the conventional staining techniques. Humans and machines can interact to solve this problem. We developed a new temporary staining protocol associated with a computer program for the diagnosis of Cryptosporidium spp. oocysts in fecal samples. We established 62 different temporary staining conditions by studying 20 experimental protocols. Cryptosporidium spp. oocysts were concentrated using the Three Fecal Test (TF-Test®) technique and confirmed by the Kinyoun method. Next, we built a bank with 299 images containing oocysts. We used segmentation in superpixels to cluster the patches in the images; then, we filtered the objects based on their typical size. Finally, we applied a convolutional neural network as a classifier. The trichrome modified by Melvin and Brooke, at a concentration use of 25%, was the most efficient dye for use in the computerized diagnosis. The algorithms of this new program showed a positive predictive value of 81.3 and 94.1% sensitivity for the detection of Cryptosporidium spp. oocysts. With the combination of the chosen staining protocol and the precision of the computational algorithm, we improved the Ova and Parasite exam (O&P) by contributing in advance toward the automated diagnosis.

Computerized analysis of seedling performance in evaluating the phytotoxicity of chemical treatment of soybean seeds

Chemical treatment of soybean seeds is very important to ensure successful crop establishment. However, problems such as phytotoxicity of product combinations that can reduce seed physiological performance require attention. The use of computational resources has shown potential in identifying phytotoxic effects and contributing to the steps of quality control of treated seeds. The aim of this study was to determine if computerized image analysis of seedlings enables the phytotoxicity of chemical treatment of soybean seeds to be assessed in an effective and simplified manner. Samples from two soybean seed lots were treated with fungicides, insecticides, micronutrients, and their combinations, as well as with polymer and drying powder (coatings). After chemical treatment, the seeds were evaluated for germination, first germination count, seedling emergence in sand, accelerated aging, and seedling performance with and without the correction of regions not automatically demarcated (Vigor-S). We found high correlation of the Vigor-S parameters with the traditional tests for detection of phytotoxic effects of chemical treatment, regardless of correction made in the system. Computerized image analysis of seedlings is an effective and highly sensitive resource for evaluating possible phytotoxicity effects due to chemical treatment of soybean seeds.

Computerized image analysis of blood vessels within mucosal defects for the prediction of delayed bleeding following colonic endoscopic mucosal resection: a pilot study

Background Clinically significant post-endoscopic bleeding (CSPEB) is a common complication following colonic endoscopic mucosal resection (EMR). Current prediction tools are clinical and do not use the appearance of the post-EMR mucosal defect. We aimed to predict CSPEB by analyzing blood vessel morphology within the post-EMR mucosal defect. Methods 43 patients with CSPEB were matched to 43 non-bleeders for clinical variables associated with CSPEB. Computerized image analysis quantified the morphologic characteristics of the blood vessels in the defect. Variables were measured in relation to the mucosal defect area. Multivariate analysis and a neural network (NNET) were used as prediction models. Results The CSPEB group vessels had larger maximum diameter (113.07 vs. 69.03; P < 0.001), larger minimum radius (5.09 vs. 3.28; P = 0.002), larger perimeter value (337.82 vs. 193.86; P < 0.001), larger vessel length-of-outline (351.83 vs. 220.68; P = 0.002), and larger fractal dimension (1.11 vs. 1.10; P = 0.005) compared with non-bleeders. Discriminant analysis yielded 86 % sensitivity and 76.7 % specificity and an NNET classifier yielded 100 % sensitivity and 76.9 % specificity for identifying patients at risk. Conclusions Blood vessel morphology in the post-EMR defect can be used to predict bleeding following colonic EMR.