Recently, most image-level supervised instance segmentation methods extend Class Attention Maps (CAMs) to find the entire instance masks. Inter-pixel Relation Network (IRNet) can effectively generate the class-wise boundary maps for attention score propagation. However, class-wise boundary is likely to cause the failure of segmentation among instances. In this work, we find instance-wise information can be extracted from the displacement field of IRNet. Motivated by the observations, an improved IRNet-based instance segmentation method with instance-wise boundary has been developed. Experimental results based on PASCAL VOC 2012 demonstrate the effectiveness of our proposed method. Compared with the recent state-of-the-art methods, the mean average precision can be increased by 4.3% without any additional annotations.
Background
Marigold (Tagetes erecta L) accounts for over half of the world’s loose flower production, and marigold crop residue (MCR) are abundantly available and should be used as a forage. In this study, MCR from the last commercial flower pickings was ensilaged with lactic acid bacteria (LAB) and the shift in their volatile organic compounds (VOCs) profiles was monitored. Samples were collected at 6 different times during ensilage (3, 6, 9, 12, 15, 30 days) to determine and quantify the VOCs changes using a solid-phase microextraction (SPME) technique and gas chromatography – mass spectrometry (GC-MS).
Results
After 30 days, the caryophyllene and piperitone, which account for 14.7 and 12.1% of total VOCs, decreased by 32.9 and 9.6% respectively, alcohols increased from 2.8 to 8.1%, and the acetic acid content increased by 560%.
Conclusion
We have confirmed LAB can degrade the content of terpenes and enhance the content of alcohols and acids in MCR, which was for the first time on terpene degradation in fodder by ensilage. These results have shed light on our understanding of how to improve fodder odor and to enhance terpene degradation by lactic acid bacteria fermentation.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.