Explicit foreground and background modeling in the classification of text blocks in scene images

Abstract

Achieving high accuracy for classifying foreground and background is an interesting challenge in the field of scene image analysis because of the wide range of illumination, complex background, and scale changes. Classifying foreground and background using bag-of-feature model gives a good result. However, the performance of the classifier depends on designed features. Therefore, this paper presents an alternative classification method based on three categories of object-attributes features namely object description, color distribution and gradient strength. Each feature is computed to a classifier model. The robustness of the method has been tested on the ICDAR2015 dataset. The experimental results show that the performance of the proposed method performs competitively against the results of existing methods in term of precision and recall.

Multi-script text versus non-text classification of regions in scene images

“…Although localization of structural features by SIFT is robust against image transformations and small geometric distortions, it sometimes provides an insufficient number of keypoints in order to obtain higher-level object attributes [29] as shown in Fig. 5.…”

“…Unlike the situation of ordinary document analysis, the BG class is a container, in which both simple signs or plate colors are represented as well as complicated natural or urban scenes. Therefore, as we have suggested elsewhere [29], explicit modeling is required: intensity differences are not a sufficient heuristic for FG/BG separation. These feature vectors should be gathered into categories.…”

“…Such local features can be computed easily and have been shown to be well suited for matching and recognition as well as for many other applications where occlusion, background clutter, and other content variations occur. To compute points of interest (POI) in the image, it has been found that the Harris-Laplace (HL) detection method provides a larger portion of salient image patches than SIFT [29]. HL-detected POIs identify rich local information allowing higher-level object attributes, e.g., descriptions and color, to be completed for that region.…”

“…1. We experimented with some feature schemes, which did not produce a promising outcome [29]. The word image at the top was taken on a sunny morning whereas the photo at the bottom was taken in the evening under cloudy conditions.…”

Keypoint-based scene-text detection and character classification using color and gradient features

“…We experimented with some feature schemes, which did not produce a promising outcome (Sriman and Schomaker 2015a). The word image at the top was taken on a sunny morning whereas the photo at the bottom was taken in the evening under cloudy conditions.…”