An Efficient CBIR System for High Resolution Remote Sensing Images

MF2C3: Multi-Feature Fuzzy Clustering to Enhance Cell Colony Detection in Automated Clonogenic Assay Evaluation

“…As a final design choice, we consistently maintained the same size for both the sliding window W ij (with ω × ω pixels) and the squared neighborhood N (x k ) (with size ν × ν pixels) employed in the spatial FCM clustering algorithm. Therefore, also supported by the literature about local feature extraction [56], we set ν = ω = 5 pixels for all the experiments. Therefore, for a balance between the intensity and spatial components, as well as being supported by [7,9], both the functioning parameters and were set to 1 in our implementation.…”

“…Therefore, for a balance between the intensity and spatial components, as well as being supported by [7,9], both the functioning parameters and were set to 1 in our implementation. Since this is the first approach based on local descriptors for clonogenic assay evaluation, we investigated the size of the sliding window , composed of × pixels, since it might affect the extracted textural information [56]. With this goal, we considered a calibration set consisting of 20 randomly selected images (5 for each cell line described in Section 3.1) and the values ∈ {3, 5, 7, 9} were tested in terms of the clonogenic assay evaluation performance, measured as the absolute error between the automatic area-based SF against the manual gold-standard measurement (i.e.,…”

“…As a final design choice, we consistently maintained the same size for both the sliding window (with × pixels) and the squared neighborhood ( ) (with size × pixels) employed in the spatial FCM clustering algorithm. Therefore, also supported by the literature about local feature extraction [56], we set = = 5 pixels for all the experiments.…”