Tipus de document
Treball de fi de grauData de publicació
Llicència de publicació
Si us plau utilitzeu sempre aquest identificador per citar o enllaçar aquest document: https://hdl.handle.net/2445/232235
Image error classification with OCR and CV in robot testing
Títol de la revista
Autors
Director/Tutor
Contribució addicional
ISSN de la revista
Títol del volum
Resum
Image classification has traditionally relied on visual features extracted from images using computer vision models. However, in many real-world scenarios, visual information alone is insufficient to accurately determine the semantic content of an image, particularly when relevant information is conveyed through embedded text. Examples such as product packaging, documents, posters, or signs often require understanding both visual appearance and textual content to achieve reliable classification.
This project presents a multimodal image classification approach that combines visual information with text extracted through Optical Character Recognition (OCR). A computer vision model is used to capture visual features, while the textual content identified by the OCR system is processed through a complementary text classification strategy. The outputs of both modalities are integrated to improve the robustness and accuracy of the final prediction, allowing each modality to compensate for the limitations of the other.
The proposed approach demonstrates that incorporating textual information can significantly enhance classification performance in cases where visual cues are ambiguous or insufficient. By leveraging complementary sources of information, the multimodal framework provides a more comprehensive understanding of image content and illustrates the benefits of combining computer vision and natural language processing techniques for image classification tasks.
Descripció
Treballs finals del Màster de Fonaments de Ciència de Dades, Facultat de matemàtiques, Universitat de Barcelona. Any: 2026. Tutor: Laura Igual Muñoz
Recurs relacionat
Matèries (anglès)
Citació
Citació
BALLESTEROS FELIPE, Iván. Image error classification with OCR and CV in robot testing. [consulted: 22 of November of 2026]. Available at: https://hdl.handle.net/2445/232235