Image Captioning on Diseased Chili Plants Using ResNeXt50 and Transformer

Chili is a popular vegetable globally, especially in Indonesia, where it is the leading producer. However, its cultivation faces challenges like pest infestations and diseases, which can disrupt yield and increase prices. To overcome this, technology was developed to assess plant conditions, focusing on using image captioning to identify and describe affected areas, especially when pests or diseases are present. Image captioning involves capturing images and generating descriptive sentences. The proposed method uses ResNeXt50 as the encoder and Transformer as the decoder for image captioning. Experimental results showed a BLEU score of 30.52%, indicating good quality in image descriptions or captions. However, there is room for improvement to achieve better results. Therefore, the combination of ResNeXt50 and Transformer for image captioning demonstrates good performance, but further enhancements are needed for optimal outcomes.