Ç¥ÁØÈ­ Âü¿©¾È³»

TTAÀÇ Ç¥ÁØÇöȲ

Ȩ > Ç¥ÁØÈ­ °³¿ä > TTAÀÇ Ç¥ÁØÇöȲ

Ç¥ÁعøÈ£ TTAK.KO-10.1498-Part2 ±¸Ç¥ÁعøÈ£
Á¦°³Á¤ÀÏ 2024-12-06 ÃÑÆäÀÌÁö 29
ÇѱÛÇ¥Áظí ÀÚÀ²ÁÖÇà AI ÇнÀ¿ë µ¥ÀÌÅÍ - Á¦2ºÎ: °´Ã¼ ºÐ·ùü°è
¿µ¹®Ç¥Áظí AI Training Data for Automated Driving System - Part 2: Object Taxonomy
Çѱ۳»¿ë¿ä¾à ÀÌ Ç¥ÁØ¿¡¼­´Â ÀÚÀ²ÁÖÇà AI ÇнÀ¿ë µ¥ÀÌÅÍ ±¸ÃàÀ» À§ÇÑ ¶óº§¸µ ´Ü°è¿¡¼­ µµ·Î»óÀÇ °´Ã¼ ºÐ·ùü°è¸¦ Á¤ÀÇÇÑ´Ù. ÀÌ Ç¥ÁØ¿¡¼­´Â ÀÚÀ²ÁÖÇà ±â´ÉÀ» À§ÇÑ ¼¾¼­¿Í µ¥ÀÌÅÍ À¶ÇÕ À¯´Ö »çÀÌÀÇ µ¥ÀÌÅÍ Åë½Å- ³í¸® ÀÎÅÍÆäÀ̽º[2]¿Í Àϰü¼ºÀ» À¯ÁöÇϱâ À§ÇÏ¿© °´Ã¼ ¼öÁØÀÇ ÀÎÅÍÆäÀ̽º¿¡ ºÎÇÕÇϵµ·Ï °´Ã¼ ºÐ·ùü°è¸¦ Á¤ÀÇÇÑ´Ù. ÀÚÀ²ÁÖÇà µµ·Îȯ°æ¿¡¼­ ÇʼöÀûÀ¸·Î ÀνÄÇØ¾ß ÇÏ´Â °´Ã¼ÀÇ Á¾·ù¸¦ µ¿Àû °´Ã¼, µµ·Î °´Ã¼, Á¤Àû °´Ã¼, ºó°ø°£ ¿µ¿ª °´Ã¼·Î ºÐ·ùÇÏ°í °´Ã¼ÀÇ »óÅÂÁ¤º¸¸¦ Á¤ÀÇÇÑ´Ù.
¶óº§¸µ Ç¥±â¿¡¼­ °´Ã¼ÀÇ Á¾·ù¸¦ ºÐ·ùÇÒ ¼ö Àִ ü°è¸¦ Á¦°øÇÏ¿© ÀÚÀ²ÁÖÇà AI ÇнÀ¿¡ ÀÌ¿ëÇÒ ¼ö ÀÖµµ·Ï µ¥ÀÌÅÍ °ü¸® ȯ°æÀ» Á¦°øÇÑ´Ù. ÁÖ¼®Àº À̹ÌÁö À̸§, µµ·Î»ó °´Ã¼ÀÇ Á¾·ù, °´Ã¼ÀÇ ÁÂÇ¥, °´Ã¼ÀÇ Å©±â µîÀ» Æ÷ÇÔÇÑ´Ù.
¿µ¹®³»¿ë¿ä¾à This standard defines a classification scheme for objects on the road in the labeling phase for building data for AI training. This standard defines an object classification system in accordance with the object-level interface to maintain consistency with the data communication and logic interface[2] between sensors and data fusion units for autonomous driving functions. It classifies the types of objects that must be recognized in an automated driving road environment as dynamic objects, road objects, and static objects, and defines the status information of the objects.
It provides a data management environment that can be used for automated driving AI learning by providing a system to classify types of objects in labeling notations. Annotations include image name, type of road object, coordinates of the object, and size of the object.
±¹Á¦Ç¥ÁØ
°ü·ÃÆÄÀÏ TTAK.KO-10.1498-Part2.pdf TTAK.KO-10.1498-Part2.pdf            

ÀÌÀü
ÁÖ¼ÒÁö½Ä¸ðµ¨ - Á¦5ºÎ: ÁÖ¼Ò ¿ë¾îÀÇ ÀǹÌÀû Ç¥Çö ¸ðµ¨
´ÙÀ½
½º¸¶Æ®°øÀå¿¡¼­ÀÇ µ¥ÀÌÅÍ °ü¸®¸¦ À§ÇÑ ºí·ÏüÀÎ - Á¦2ºÎ: ¿ä±¸»çÇ×