Text embedded in images is one of many important cues for indexing and retrieval of images and videos. In the paper, we present a novel method of detecting text aligned either horizontally or vertically, in which a py...Text embedded in images is one of many important cues for indexing and retrieval of images and videos. In the paper, we present a novel method of detecting text aligned either horizontally or vertically, in which a pyramid structure is used to represent an image and the features of the text are extracted using SUSAN edge detector. Text regions at each level of the pyramid are identified according to the autocorrelation analysis. New techniques are introduced to split the text regions into basic ones and merge them into text lines. By evaluating the method on a set of images, we obtain a very good performance of text detection.展开更多
This paper proposes a learning-based method for text detection and text segmentation in natural scene images. First, the input image is decomposed into multiple connected-components (CCs) by Niblack clustering algorit...This paper proposes a learning-based method for text detection and text segmentation in natural scene images. First, the input image is decomposed into multiple connected-components (CCs) by Niblack clustering algorithm. Then all the CCs including text CCs and non-text CCs are verified on their text features by a 2-stage classification module, where most non-text CCs are discarded by an attentional cascade classifier and remaining CCs are further verified by an SVM. All the accepted CCs are output to result in text only binary image. Experiments with many images in different scenes showed satisfactory performance of our proposed method.展开更多
Although investigating metaphors in advertising is gaining in popularity, there are still certain unresolved arguments, such as the interaction between elements of different modalities. This study,composed of three be...Although investigating metaphors in advertising is gaining in popularity, there are still certain unresolved arguments, such as the interaction between elements of different modalities. This study,composed of three behavioral experiments, aims to identify how verbal anchoring(literal anchoring, metaphor anchoring and unrelated anchoring) influences the processing of pictorial metaphors in advertising, by observing the cognitive and affective indicators, advertising comprehension and advertising likeability. The results showed 1) that metaphors in pictorial modality were recognized more quickly than those in verbal modality, 2) that verbal anchoring facilitated participants ’ comprehending and appreciating of pictorial metaphors and 3) that literally-anchored metaphors with a moderate level of novelty yielded the most favorable cognitive responses. The study not only enriches the existing theoretical framework of multimodal metaphors in advertising, but also proposes an optimal match between pictorial metaphors and verbal elements, for advertisers and manufacturers to design effective multimodal advertisements.展开更多
文摘Text embedded in images is one of many important cues for indexing and retrieval of images and videos. In the paper, we present a novel method of detecting text aligned either horizontally or vertically, in which a pyramid structure is used to represent an image and the features of the text are extracted using SUSAN edge detector. Text regions at each level of the pyramid are identified according to the autocorrelation analysis. New techniques are introduced to split the text regions into basic ones and merge them into text lines. By evaluating the method on a set of images, we obtain a very good performance of text detection.
基金Project supported by the OMRON and SJTU Collaborative Founda-tion under PVS project (2005.03~2005.10)
文摘This paper proposes a learning-based method for text detection and text segmentation in natural scene images. First, the input image is decomposed into multiple connected-components (CCs) by Niblack clustering algorithm. Then all the CCs including text CCs and non-text CCs are verified on their text features by a 2-stage classification module, where most non-text CCs are discarded by an attentional cascade classifier and remaining CCs are further verified by an SVM. All the accepted CCs are output to result in text only binary image. Experiments with many images in different scenes showed satisfactory performance of our proposed method.
基金supported by the National Social Science Foundation of China (Grant No. 19BYY088)。
文摘Although investigating metaphors in advertising is gaining in popularity, there are still certain unresolved arguments, such as the interaction between elements of different modalities. This study,composed of three behavioral experiments, aims to identify how verbal anchoring(literal anchoring, metaphor anchoring and unrelated anchoring) influences the processing of pictorial metaphors in advertising, by observing the cognitive and affective indicators, advertising comprehension and advertising likeability. The results showed 1) that metaphors in pictorial modality were recognized more quickly than those in verbal modality, 2) that verbal anchoring facilitated participants ’ comprehending and appreciating of pictorial metaphors and 3) that literally-anchored metaphors with a moderate level of novelty yielded the most favorable cognitive responses. The study not only enriches the existing theoretical framework of multimodal metaphors in advertising, but also proposes an optimal match between pictorial metaphors and verbal elements, for advertisers and manufacturers to design effective multimodal advertisements.
