YOLO stuff in perception#8
Conversation
| [PLACEHOLDER] & [PLACEHOLDER] & [PLACEHOLDER] \\ | ||
| \hline | ||
| \end{tabular} | ||
| \end{table} |
There was a problem hiding this comment.
this table is overflowing on the page and just has placeholders, this needs to be fixed.
There was a problem hiding this comment.
also, adding maybe an interpretation of the results once you have them would be good
| drops out once it leaves the field of view and reappears on the first | ||
| frame it is detected again. | ||
|
|
||
| \paragraph{Performance.} \textit{we lwk did not take performance metrics, but will add them later if we have time} |
There was a problem hiding this comment.
need to comment this out
| annotations easy to spot. We then fine-tuned a YOLO11n detector on each | ||
| dataset. | ||
|
|
||
| \paragraph{Dataset merge.} An automated pipeline merges the five |
There was a problem hiding this comment.
more detail on the parameters of the pipeline would be nice
There was a problem hiding this comment.
i.e., what network size and why, activation functions? experimentations? additional stuff?
| YOLO11 model is then trained on the merged dataset, and those weights | ||
| are the detector we deploy on the vehicle. | ||
|
|
||
| \paragraph{ZED ROS 2 node.} On the vehicle, the merged model runs |
There was a problem hiding this comment.
a bit conversational, more mathematical rigor here would be nice.
| [PLACEHOLDER] & [PLACEHOLDER] & [PLACEHOLDER] \\ | ||
| \hline | ||
| \end{tabular} | ||
| \end{table} |
There was a problem hiding this comment.
also, adding maybe an interpretation of the results once you have them would be good
2 participants
No description provided.