Privately Owned Vehicle Work Group Meeting - 2025/01/27 - Slot 2

[m-zain-khawaja]

Agenda

• Update on AutoSeg network diagram with new names
• Update on Scene3D (previously SuperDepth) training
• Update on EgoPath (previously PathDet) dataset curation
• Update on EgoLanes (previously LaneDet) dataset curation

Discussion

Scene3D Network Training Update

@m-zain-khawaja :

Second phase of Scene3D training using both simulator and real world data is progressing well. Mutiple experiments have been conducted to establish the best balance in the loss function and cross-dataset loss weighting. The DDAD, UrbanSyn and KITTI datasets are being used for training and the Argoverse Dataset is being used as a large and diverse test set. Full results with visualizations should be ready for sharing ahead of next week's WG meeting. Only testing and visualization scripts need to be written, the goal is to complete these scripts this week - on track with plan.

EgoPath Dataset Curation Update

@siddas27 has completed dataset curation for the Comma2k19 dataset and uploaded to Kaggle - his PR has been merged with the main branch - many thanks.

@TranHuuNhatHuy expects to hopefully finish CurveLanes dataset curation this week.

@sarun-hub is currently performing data auditing for the ROADWorks dataset, it is also expected to be completed this week.

@docjag has submitted the latest fixes in his PR - this is also expected to be completed this week.

Overall, dataset parsing should be completed on track with plan.

Dataset curation tracking

• TuSimple - 6,394 training data samples
• CULane - 15,800 training data samples
• CurveLane - (parsing completed by @TranHuuNhatHuy, data to be uploaded soon)
• BDD100K - (parsing completed by @sarun-hub, data to be uploaded soon)
• ROADWorks - (parsing completed by @docjag, data to be uploaded soon)
• Comma2k19 - 31,962 training data samples

EgoLanes Dataset Curation Update

• TuSimple - 6,394 training data samples
• CULane - 15,800 training data samples
• CurveLane - (to be started)
• BDD100K - (to be started)

@devang-marvania has completed parsing the CULane dataset for the EgoLanes neural network and has uploaded it to Kaggle

Attendees

• @m-zain-khawaja (Senior Tech Lead at AWF and Privately Owned Vehicle Work Group Chair)
• Christian John (President at TierIV North America and AWF Strategic Planning Committee Chair)
• Samet Kutuk (Marketing Steering Committee Chair at AWF)
• Mohammad Musa (CEO at DeepenAI and AWF Safety Assurance Work Group Lead)
• Armagan Arslan (CEO at LEODrive and AWF Strategic Planning Committee Vice Chair)
• Sergey Christ (CEO at Alternative Embedded Innovation)
• David Walmroth (Co-Chair of OpenADKit Work Group at AWF and US Director of Business and Project Development at PIX Moving)
• Aadarsh Tiwari (Software Engineer at KPIT)
• Atanasko Mitrev (Innovation R&D at EXO Service Solutions)
• Mahmoud Dahmani (Robotics Software Engineer at Vecna Robotics)
• Augustine Osarogiagbon (Canada Revenue Agency, Information Technology Analyst/Developer)
• Alexander Korvyakov (Senior Product Manager - MOIA by Volkswagen)

Zoom Meeting Video Recording

Video Meeting Link
Please contact the work group lead (@m-zain-khawaja) to request access to a recording of this meeting.