Changes between Initial Version and Version 1 of Hardware/Cameras


Ignore:
Timestamp:
Mar 23, 2022, 1:04:12 AM (3 years ago)
Author:
mg4089
Comment:

—

Legend:

Unmodified
Added
Removed
Modified

  • Hardware/Cameras

    v1 v1  
     1
     2== Installed Camera Information
     3
     4**Hikvision DS-2CD5585G0-IZHS**
     5    Mudd floor 1, towards 120th street
     6    Mudd first floor Mech E lab
     7    [https://www.hikvision.com/uploadfile/image/11769_DatasheetofDS2CD5585G0IZ(H)S.pdf]
     8
     9**Hikvision DS-2CD5585G0-IZHS**
     10    Mudd 2nd floor, towards Amsterdam balcony camera
     11
     12**Hikvision DS-2CD4AC5F-IZH **
     13    12MP Outdoor Network Bullet Camera with Night Vision, 2.8-12mm Lens & Built-In Heater S/N: 212528616:
     14    20191015: installation in Botwinick lab mudd1224, towards 120th street
     15**Hikvision DS-2CD5585G0-IZHS  **
     16    2.8mm-12mm (8MP)
     17    Mudd 12thFloor Balcony towards Amsterdam Avenue
     18
     19
     20== COSMOS Cameras Data-set
     21**1st-floor videos (anonymized*)**: [https://drive.google.com/drive/u/0/folders/1QXrfsLXEKfRfQyc6qzvtg37A0Z1i0io5]
     22
     23**2nd-floor videos (anonymized*)**: [https://drive.google.com/drive/u/0/folders/1LR7H4theRazz2_uYHvCFGVVewQmKbWSF]
     24
     25**12th-floor videos (120th street)**: [https://drive.google.com/drive/u/0/folders/1SEsocAAIReepdjE4XyVyT4kiqrunv7BU]
     26
     27**12th-floor videos (Amsterdam Avenue)**: [https://drive.google.com/drive/u/0/folders/1qC-62s8ohTGg-odyzo7BNw2GDv1OIeoK]
     28
     29Videos in this directory are outputs of the COSMOS YOLOv4 blurring pipeline.
     30
     31Faces and license plates are anonymized with Gaussian blurred areas defined by bounding box detection coordinates. A high-level overview of the blurring.
     32
     33***Anonymization workflow:**
     34
     351. Frames are read individually from a video file.
     36
     372. Each frame is then:
     38
     39    2.1 Resized to the input size of the specific YOLOv4 model (960x960 or 1440x1440)
     40
     41    2.2 Passed through the YOLOv4 model which outputs bounding box predictions
     42
     43    2.3 Blurred corresponding to the bounding box predictions
     44
     453. Blurred frames are written to the output video file.
     46
     47The YOLOv4 blurring model is trained in Darknet on the Mudd 1st floor video
     48dataset annotated in Summer 2021. Models are converted from Darknet to PyTorch for
     49integration into the current implementation of the blurring pipeline.