MY196904A - System and method for anomaly detection and profiling of extracted raw image for processing & value analysis - Google Patents

Abstract

The present invention relates to a system and method for anomalies detection by utilizing image processing technique and calculation of materials atomic number. The system of the present invention comprises of two (2) main modules: an object structure elimination module (104); and a suspicious object detection module (106). The object structure elimination module (104) further comprises of X-ray image extraction units (202, 206, 208) for extracting unique features of object image; an outer boundary locator unit (210) for identifying outer contour of the object image; an inner frame structure locater unit (212) or identifying inner structure of the object image; an object structure and inner content separation unit (214) for distinguishing the outer contour and the inner structure of the object image. The suspicious object detection module further comprises of: an object content segmentation module (304) for performing segmentation of the anomalies region whereby the anomalies region having high density and intensity level is identified based on the density obtained from the dual energy X-Ray output in accordance to predefined rules; and a machine learning classification unit (306) for analysing the said anomalies region to determine probability of existence of anomaly based on machine learning algorithm (306), classifier model (312) and a minimum predetermined threshold. The most illustrative drawing is FIG. 2.0.