HIPPOCAMPAL · TV2V · WORKFLOW 3.5 · GATE B

Task_198
dashcam_video_super_resolution

Restore a low-resolution 320x180 dashcam interval to its aligned sharp 1280x720 source-backed Target. Samples use balanced windows near 81 frames (76–87 in the full census), preserving every source frame exactly once.

public review evidence. Synchronized mode: Start, pause, and seek keep Input and Target aligned. Owner must complete this page before requesting the Human Evaluator. Both people answer the same evidence-bound PASS/FAIL question.

8validation data units
32four-file objects
15,889,151payload bytes
8review samples
17declared strata
Repository / PRhipo-dev/tv2v-reasoning · #565 Human EvaluatorSkyeliForWork Producing commitc078a850f7f4c823d6057c7a6ec8a8298a87e51d Full generation commitnot applicable at Gate B Full generation task rootnot applicable at Gate B Full generation treenot applicable at Gate B Pipelinedashcam_video_super_resolution · 0.2.0 Generation run918dd43471e2d04bbbd2158a1d147cc6a1c2d456cd7d101362cdca70004e4bbf Run manifestcc259cb937ea907b0e5cc343d5546a029b678faf5f0b21c744be0acb7b7fa0a3 Canonical data identitys3://tv2v-data/validation/Task_198_dashcam_video_super_resolution/runs/918dd43471e2d04bbbd2158a1d147cc6a1c2d456cd7d101362cdca70004e4bbf/gate-b/ Review evidence identitys3://tv2v-data/validation/Task_198_dashcam_video_super_resolution/runs/918dd43471e2d04bbbd2158a1d147cc6a1c2d456cd7d101362cdca70004e4bbf/gate-b/ Dataset identitynot applicable at Gate B Evidence digest1c46cb05a8222c048c56a5f073dc42d35930cc20432fd250de35c498670e75de Generated2026-08-13T03:59:53.255388+00:00

Reviewer role

The page stores each role separately in this browser. Owner completion must be imported from the official Owner attestation before the Human Evaluator column becomes editable.

Machine result

CheckStatusSummary
repositoryPASSExact producing commit and single Task 198 tree are bound.
task-testsPASSTask window, Metadata, synthetic partition, and real-source smoke tests passed.
four-file-contractPASSAll eight samples contain exactly four canonical non-empty files.
metadata-1.2PASSAll eight metadata records pass Metadata 1.2 and bind the exact source window.
pipeline-commitPASSEvery sample binds pipeline 0.2.0 at the clean exact producing commit.
near-81-policyPASSEvery source is partitioned into contiguous balanced windows nearest 81 frames; every source frame is covered once.
source-pairingPASSInput and Target derive from the same exact source-backed interval.
task-semanticsPASSInput is 320x180 and Target is aligned 1280x720, preserving scene, motion, timing, colour, and content.
mediaPASSAll 16 videos completely decode with equal per-pair frame count and presentation timeline.
determinism-resume-no-overwritePASSResume skipped 8/8 with zero Raw member bytes; fresh generation matched 32/32 files byte-for-byte.
validation-s3-bindingPASSAll 32 payloads match current VersionIds and full exact-version SHA-256 readback.
coveragePASSSelection covers train/val/test, 76/87 boundaries, 81-frame core, source beginnings/endings, rotations 0/90/270, and rare yuvj420p.
range-sourcePASSValidation used pinned exact-VersionId ZIP ranges and no full archive download.

Coverage

Expected full units: 43,806 · The full 3,000-source census partitions all 3,524,537 decoded frames into 43,806 contiguous non-overlapping balanced windows nearest 81 frames, with no exclusions, dropped frames, or duplicates.

StratumCountComplete sample IDs
76-frame-minimum1Task_198_dashcam_video_super_resolution_000000677
78-frame1Task_198_dashcam_video_super_resolution_000043805
80-frame1Task_198_dashcam_video_super_resolution_000014654
81-frame4Task_198_dashcam_video_super_resolution_000007088, Task_198_dashcam_video_super_resolution_000008445, Task_198_dashcam_video_super_resolution_000011458, Task_198_dashcam_video_super_resolution_000029195
87-frame-maximum1Task_198_dashcam_video_super_resolution_000000240
bright-stratum1Task_198_dashcam_video_super_resolution_000008445
dark-stratum1Task_198_dashcam_video_super_resolution_000011458
final-ordinal1Task_198_dashcam_video_super_resolution_000043805
rare-yuvj420p1Task_198_dashcam_video_super_resolution_000007088
rotation-01Task_198_dashcam_video_super_resolution_000011458
rotation-2703Task_198_dashcam_video_super_resolution_000000677, Task_198_dashcam_video_super_resolution_000008445, Task_198_dashcam_video_super_resolution_000043805
rotation-904Task_198_dashcam_video_super_resolution_000000240, Task_198_dashcam_video_super_resolution_000007088, Task_198_dashcam_video_super_resolution_000014654, Task_198_dashcam_video_super_resolution_000029195
source-beginning6Task_198_dashcam_video_super_resolution_000000240, Task_198_dashcam_video_super_resolution_000007088, Task_198_dashcam_video_super_resolution_000008445, Task_198_dashcam_video_super_resolution_000011458, Task_198_dashcam_video_super_resolution_000014654, Task_198_dashcam_video_super_resolution_000029195
source-ending2Task_198_dashcam_video_super_resolution_000000677, Task_198_dashcam_video_super_resolution_000043805
test_002Task_198_dashcam_video_super_resolution_000029195, Task_198_dashcam_video_super_resolution_000043805
train_005Task_198_dashcam_video_super_resolution_000000240, Task_198_dashcam_video_super_resolution_000000677, Task_198_dashcam_video_super_resolution_000007088, Task_198_dashcam_video_super_resolution_000008445, Task_198_dashcam_video_super_resolution_000011458
val_001Task_198_dashcam_video_super_resolution_000014654
SAMPLE 01

Task_198_dashcam_video_super_resolution_000000240

train_0087-frame-maximumsource-beginningrotation-90
not fully played

Source: BDD100K via SpicyqSama007/nltk_data · 00091078-84635cf2/frames-000000-000086
Selection: Maximum near-81 boundary at the start of a rotated train source.

Input and Target not completed.
INPUT · 87 frames · 320×180 · 30.0 fps
0:00 / --:--
TARGET · 87 frames · 1280×720 · 30.0 fps
0:00 / --:--
EXACT PROMPTRestore this low-resolution dashcam video to a sharp 1280x720 video while preserving the original scene, colors, motion, timing, and all visible content.
Watch for
  1. Watch the complete synchronized pair; confirm Input and Target have identical motion, framing, timing, and visible content.
  2. Confirm Target is genuinely sharper at 1280x720 while Input is the faithful 320x180 area downsample.
  3. Confirm smooth complete playback for this 87-frame near-81 window, including its first and last frame.
Complete exact metadata.json
{
  "schema_version": "1.2",
  "sample_id": "Task_198_dashcam_video_super_resolution_000000240",
  "task": {
    "id": "Task_198",
    "name": "dashcam_video_super_resolution",
    "modality": "v2v"
  },
  "source": {
    "dataset": "BDD100K via SpicyqSama007/nltk_data",
    "split": "train_00",
    "sample_id": "00091078-84635cf2/frames-000000-000086",
    "uri": "s3://tv2v-data/raw/taxonomy/Task_198_Dashcam_Video_Super_Resolution/SpicyqSama007_nltk_data_bdd100k_3k/archives/train_00/bdd100k_videos_train_00.zip"
  },
  "media": {
    "input": {
      "file": "input_video.mp4",
      "width": 320,
      "height": 180,
      "fps": 30.0,
      "num_frames": 87
    },
    "target": {
      "file": "target_video.mp4",
      "width": 1280,
      "height": 720,
      "fps": 30.0,
      "num_frames": 87
    }
  },
  "pipeline": {
    "name": "dashcam_video_super_resolution",
    "version": "0.2.0",
    "commit": "c078a850f7f4c823d6057c7a6ec8a8298a87e51d"
  },
  "optional": {
    "source_member": "bdd100k/videos/train/00091078-84635cf2.mov",
    "source_sha256": "ddc890d9847cf2d793f6dee787db913fab8df445da26770fc462dd38ec8fb9b2",
    "source_decoded_frames": 606,
    "source_r_frame_rate": "60000/1001",
    "source_rotation_degrees": 90,
    "source_pixel_format": "yuv420p",
    "source_encoded_width": 720,
    "source_encoded_height": 1280,
    "window": {
      "ordinal": 0,
      "frame_start_inclusive": 0,
      "frame_end_exclusive": 87,
      "num_frames": 87,
      "target_frames": 81,
      "partition_policy": "nearest-window-count-balanced-full-coverage-v1"
    }
  },
  "identity": {
    "sample_fingerprint": "bcfba588d01bbb64e3e5f4c02860c377c5ddaf95c32a69c90bdb6f4b3385ac20"
  }
}
Per-sample automated checks
CheckStatusSummary
task198.super-resolution-pairPASSInput is a faithful 320x180 area downsample of the aligned 1280x720 source-backed Target.

Required human checks

Exactly three checks are required for this sample. The active human may check only their own column.

QuestionOwnerHuman Evaluator
Did you completely watch this sample's Input and Target in the required playback mode?
Does the exact Prompt describe the intended task and the visible Input-to-Target relationship?
Are the metadata, automated result, and sample-specific viewing cues consistent with the media?
SAMPLE 02

Task_198_dashcam_video_super_resolution_000000677

train_0076-frame-minimumsource-endingrotation-270
not fully played

Source: BDD100K via SpicyqSama007/nltk_data · 0024b742-acbed4fb/frames-000462-000537
Selection: Minimum near-81 boundary at the end of a rotated train source.

Input and Target not completed.
INPUT · 76 frames · 320×180 · 30.0 fps
0:00 / --:--
TARGET · 76 frames · 1280×720 · 30.0 fps
0:00 / --:--
EXACT PROMPTRestore this low-resolution dashcam video to a sharp 1280x720 video while preserving the original scene, colors, motion, timing, and all visible content.
Watch for
  1. Watch the complete synchronized pair; confirm Input and Target have identical motion, framing, timing, and visible content.
  2. Confirm Target is genuinely sharper at 1280x720 while Input is the faithful 320x180 area downsample.
  3. Confirm smooth complete playback for this 76-frame near-81 window, including its first and last frame.
Complete exact metadata.json
{
  "schema_version": "1.2",
  "sample_id": "Task_198_dashcam_video_super_resolution_000000677",
  "task": {
    "id": "Task_198",
    "name": "dashcam_video_super_resolution",
    "modality": "v2v"
  },
  "source": {
    "dataset": "BDD100K via SpicyqSama007/nltk_data",
    "split": "train_00",
    "sample_id": "0024b742-acbed4fb/frames-000462-000537",
    "uri": "s3://tv2v-data/raw/taxonomy/Task_198_Dashcam_Video_Super_Resolution/SpicyqSama007_nltk_data_bdd100k_3k/archives/train_00/bdd100k_videos_train_00.zip"
  },
  "media": {
    "input": {
      "file": "input_video.mp4",
      "width": 320,
      "height": 180,
      "fps": 30.0,
      "num_frames": 76
    },
    "target": {
      "file": "target_video.mp4",
      "width": 1280,
      "height": 720,
      "fps": 30.0,
      "num_frames": 76
    }
  },
  "pipeline": {
    "name": "dashcam_video_super_resolution",
    "version": "0.2.0",
    "commit": "c078a850f7f4c823d6057c7a6ec8a8298a87e51d"
  },
  "optional": {
    "source_member": "bdd100k/videos/train/0024b742-acbed4fb.mov",
    "source_sha256": "67aa225c9f387bf5f2f3ed54e7ae3154a4f907b964292d5aae88071470f3912b",
    "source_decoded_frames": 614,
    "source_r_frame_rate": "60000/1001",
    "source_rotation_degrees": 270,
    "source_pixel_format": "yuv420p",
    "source_encoded_width": 720,
    "source_encoded_height": 1280,
    "window": {
      "ordinal": 6,
      "frame_start_inclusive": 462,
      "frame_end_exclusive": 538,
      "num_frames": 76,
      "target_frames": 81,
      "partition_policy": "nearest-window-count-balanced-full-coverage-v1"
    }
  },
  "identity": {
    "sample_fingerprint": "764ad2199e921ccf0f112f8eb4741393b4a178ef7ac830cf731694c8aac18bc5"
  }
}
Per-sample automated checks
CheckStatusSummary
task198.super-resolution-pairPASSInput is a faithful 320x180 area downsample of the aligned 1280x720 source-backed Target.

Required human checks

Exactly three checks are required for this sample. The active human may check only their own column.

QuestionOwnerHuman Evaluator
Did you completely watch this sample's Input and Target in the required playback mode?
Does the exact Prompt describe the intended task and the visible Input-to-Target relationship?
Are the metadata, automated result, and sample-specific viewing cues consistent with the media?
SAMPLE 03

Task_198_dashcam_video_super_resolution_000007088

train_0081-framesource-beginningrare-yuvj420protation-90
not fully played

Source: BDD100K via SpicyqSama007/nltk_data · 01460ec4-0233fef9/frames-000000-000080
Selection: Exact-81 window covers the rare yuvj420p colour-range risk.

Input and Target not completed.
INPUT · 81 frames · 320×180 · 30.0 fps
0:00 / --:--
TARGET · 81 frames · 1280×720 · 30.0 fps
0:00 / --:--
EXACT PROMPTRestore this low-resolution dashcam video to a sharp 1280x720 video while preserving the original scene, colors, motion, timing, and all visible content.
Watch for
  1. Watch the complete synchronized pair; confirm Input and Target have identical motion, framing, timing, and visible content.
  2. Confirm Target is genuinely sharper at 1280x720 while Input is the faithful 320x180 area downsample.
  3. Confirm smooth complete playback for this 81-frame near-81 window, including its first and last frame.
Complete exact metadata.json
{
  "schema_version": "1.2",
  "sample_id": "Task_198_dashcam_video_super_resolution_000007088",
  "task": {
    "id": "Task_198",
    "name": "dashcam_video_super_resolution",
    "modality": "v2v"
  },
  "source": {
    "dataset": "BDD100K via SpicyqSama007/nltk_data",
    "split": "train_00",
    "sample_id": "01460ec4-0233fef9/frames-000000-000080",
    "uri": "s3://tv2v-data/raw/taxonomy/Task_198_Dashcam_Video_Super_Resolution/SpicyqSama007_nltk_data_bdd100k_3k/archives/train_00/bdd100k_videos_train_00.zip"
  },
  "media": {
    "input": {
      "file": "input_video.mp4",
      "width": 320,
      "height": 180,
      "fps": 30.0,
      "num_frames": 81
    },
    "target": {
      "file": "target_video.mp4",
      "width": 1280,
      "height": 720,
      "fps": 30.0,
      "num_frames": 81
    }
  },
  "pipeline": {
    "name": "dashcam_video_super_resolution",
    "version": "0.2.0",
    "commit": "c078a850f7f4c823d6057c7a6ec8a8298a87e51d"
  },
  "optional": {
    "source_member": "bdd100k/videos/train/01460ec4-0233fef9.mov",
    "source_sha256": "1e17f7784a490dd65d92d242b8992499be42162f1b34654ae29aeb9e1f4099df",
    "source_decoded_frames": 1208,
    "source_r_frame_rate": "19200/1",
    "source_rotation_degrees": 90,
    "source_pixel_format": "yuvj420p",
    "source_encoded_width": 720,
    "source_encoded_height": 1280,
    "window": {
      "ordinal": 0,
      "frame_start_inclusive": 0,
      "frame_end_exclusive": 81,
      "num_frames": 81,
      "target_frames": 81,
      "partition_policy": "nearest-window-count-balanced-full-coverage-v1"
    }
  },
  "identity": {
    "sample_fingerprint": "04fe2956acb65169ca08c1c7417988333a439b27da30da10e1c5fdadf48d8b78"
  }
}
Per-sample automated checks
CheckStatusSummary
task198.super-resolution-pairPASSInput is a faithful 320x180 area downsample of the aligned 1280x720 source-backed Target.

Required human checks

Exactly three checks are required for this sample. The active human may check only their own column.

QuestionOwnerHuman Evaluator
Did you completely watch this sample's Input and Target in the required playback mode?
Does the exact Prompt describe the intended task and the visible Input-to-Target relationship?
Are the metadata, automated result, and sample-specific viewing cues consistent with the media?
SAMPLE 04

Task_198_dashcam_video_super_resolution_000008445

train_0081-framesource-beginningrotation-270bright-stratum
not fully played

Source: BDD100K via SpicyqSama007/nltk_data · 019588c0-90390013/frames-000000-000080
Selection: Bright exact-81 case covers 270-degree source orientation.

Input and Target not completed.
INPUT · 81 frames · 320×180 · 30.0 fps
0:00 / --:--
TARGET · 81 frames · 1280×720 · 30.0 fps
0:00 / --:--
EXACT PROMPTRestore this low-resolution dashcam video to a sharp 1280x720 video while preserving the original scene, colors, motion, timing, and all visible content.
Watch for
  1. Watch the complete synchronized pair; confirm Input and Target have identical motion, framing, timing, and visible content.
  2. Confirm Target is genuinely sharper at 1280x720 while Input is the faithful 320x180 area downsample.
  3. Confirm smooth complete playback for this 81-frame near-81 window, including its first and last frame.
Complete exact metadata.json
{
  "schema_version": "1.2",
  "sample_id": "Task_198_dashcam_video_super_resolution_000008445",
  "task": {
    "id": "Task_198",
    "name": "dashcam_video_super_resolution",
    "modality": "v2v"
  },
  "source": {
    "dataset": "BDD100K via SpicyqSama007/nltk_data",
    "split": "train_00",
    "sample_id": "019588c0-90390013/frames-000000-000080",
    "uri": "s3://tv2v-data/raw/taxonomy/Task_198_Dashcam_Video_Super_Resolution/SpicyqSama007_nltk_data_bdd100k_3k/archives/train_00/bdd100k_videos_train_00.zip"
  },
  "media": {
    "input": {
      "file": "input_video.mp4",
      "width": 320,
      "height": 180,
      "fps": 30.0,
      "num_frames": 81
    },
    "target": {
      "file": "target_video.mp4",
      "width": 1280,
      "height": 720,
      "fps": 30.0,
      "num_frames": 81
    }
  },
  "pipeline": {
    "name": "dashcam_video_super_resolution",
    "version": "0.2.0",
    "commit": "c078a850f7f4c823d6057c7a6ec8a8298a87e51d"
  },
  "optional": {
    "source_member": "bdd100k/videos/train/019588c0-90390013.mov",
    "source_sha256": "734278d7f116edf87a61c7396723084a8fdab15ef986cca4708a67a90edf806a",
    "source_decoded_frames": 1204,
    "source_r_frame_rate": "30/1",
    "source_rotation_degrees": 270,
    "source_pixel_format": "yuv420p",
    "source_encoded_width": 720,
    "source_encoded_height": 1280,
    "window": {
      "ordinal": 0,
      "frame_start_inclusive": 0,
      "frame_end_exclusive": 81,
      "num_frames": 81,
      "target_frames": 81,
      "partition_policy": "nearest-window-count-balanced-full-coverage-v1"
    }
  },
  "identity": {
    "sample_fingerprint": "ba22d4674864dc6f6ff48af8e404a15907b09f4251ac83d3a89c93336e182c4a"
  }
}
Per-sample automated checks
CheckStatusSummary
task198.super-resolution-pairPASSInput is a faithful 320x180 area downsample of the aligned 1280x720 source-backed Target.

Required human checks

Exactly three checks are required for this sample. The active human may check only their own column.

QuestionOwnerHuman Evaluator
Did you completely watch this sample's Input and Target in the required playback mode?
Does the exact Prompt describe the intended task and the visible Input-to-Target relationship?
Are the metadata, automated result, and sample-specific viewing cues consistent with the media?
SAMPLE 05

Task_198_dashcam_video_super_resolution_000011458

train_0081-framesource-beginningrotation-0dark-stratum
not fully played

Source: BDD100K via SpicyqSama007/nltk_data · 02205e0f-20f9ad20/frames-000000-000080
Selection: Native-orientation exact-81 case separates scale handling from rotation.

Input and Target not completed.
INPUT · 81 frames · 320×180 · 30.0 fps
0:00 / --:--
TARGET · 81 frames · 1280×720 · 30.0 fps
0:00 / --:--
EXACT PROMPTRestore this low-resolution dashcam video to a sharp 1280x720 video while preserving the original scene, colors, motion, timing, and all visible content.
Watch for
  1. Watch the complete synchronized pair; confirm Input and Target have identical motion, framing, timing, and visible content.
  2. Confirm Target is genuinely sharper at 1280x720 while Input is the faithful 320x180 area downsample.
  3. Confirm smooth complete playback for this 81-frame near-81 window, including its first and last frame.
Complete exact metadata.json
{
  "schema_version": "1.2",
  "sample_id": "Task_198_dashcam_video_super_resolution_000011458",
  "task": {
    "id": "Task_198",
    "name": "dashcam_video_super_resolution",
    "modality": "v2v"
  },
  "source": {
    "dataset": "BDD100K via SpicyqSama007/nltk_data",
    "split": "train_00",
    "sample_id": "02205e0f-20f9ad20/frames-000000-000080",
    "uri": "s3://tv2v-data/raw/taxonomy/Task_198_Dashcam_Video_Super_Resolution/SpicyqSama007_nltk_data_bdd100k_3k/archives/train_00/bdd100k_videos_train_00.zip"
  },
  "media": {
    "input": {
      "file": "input_video.mp4",
      "width": 320,
      "height": 180,
      "fps": 30.0,
      "num_frames": 81
    },
    "target": {
      "file": "target_video.mp4",
      "width": 1280,
      "height": 720,
      "fps": 30.0,
      "num_frames": 81
    }
  },
  "pipeline": {
    "name": "dashcam_video_super_resolution",
    "version": "0.2.0",
    "commit": "c078a850f7f4c823d6057c7a6ec8a8298a87e51d"
  },
  "optional": {
    "source_member": "bdd100k/videos/train/02205e0f-20f9ad20.mov",
    "source_sha256": "186f9528f33f8883dbc1e9d915880bd4f9b64711bbfaf999637688b9f3851340",
    "source_decoded_frames": 1205,
    "source_r_frame_rate": "30000/1001",
    "source_rotation_degrees": 0,
    "source_pixel_format": "yuv420p",
    "source_encoded_width": 1280,
    "source_encoded_height": 720,
    "window": {
      "ordinal": 0,
      "frame_start_inclusive": 0,
      "frame_end_exclusive": 81,
      "num_frames": 81,
      "target_frames": 81,
      "partition_policy": "nearest-window-count-balanced-full-coverage-v1"
    }
  },
  "identity": {
    "sample_fingerprint": "e93b102dafe3d8477bbef6f3b1d4e39ac6531bbc4d66cd03fbb1a87196705680"
  }
}
Per-sample automated checks
CheckStatusSummary
task198.super-resolution-pairPASSInput is a faithful 320x180 area downsample of the aligned 1280x720 source-backed Target.

Required human checks

Exactly three checks are required for this sample. The active human may check only their own column.

QuestionOwnerHuman Evaluator
Did you completely watch this sample's Input and Target in the required playback mode?
Does the exact Prompt describe the intended task and the visible Input-to-Target relationship?
Are the metadata, automated result, and sample-specific viewing cues consistent with the media?
SAMPLE 06

Task_198_dashcam_video_super_resolution_000014654

val_0080-framesource-beginningrotation-90
not fully played

Source: BDD100K via SpicyqSama007/nltk_data · b1c66a42-6f7d68ca/frames-000000-000079
Selection: Validation-split near-81 case checks cross-split behaviour.

Input and Target not completed.
INPUT · 80 frames · 320×180 · 30.0 fps
0:00 / --:--
TARGET · 80 frames · 1280×720 · 30.0 fps
0:00 / --:--
EXACT PROMPTRestore this low-resolution dashcam video to a sharp 1280x720 video while preserving the original scene, colors, motion, timing, and all visible content.
Watch for
  1. Watch the complete synchronized pair; confirm Input and Target have identical motion, framing, timing, and visible content.
  2. Confirm Target is genuinely sharper at 1280x720 while Input is the faithful 320x180 area downsample.
  3. Confirm smooth complete playback for this 80-frame near-81 window, including its first and last frame.
Complete exact metadata.json
{
  "schema_version": "1.2",
  "sample_id": "Task_198_dashcam_video_super_resolution_000014654",
  "task": {
    "id": "Task_198",
    "name": "dashcam_video_super_resolution",
    "modality": "v2v"
  },
  "source": {
    "dataset": "BDD100K via SpicyqSama007/nltk_data",
    "split": "val_00",
    "sample_id": "b1c66a42-6f7d68ca/frames-000000-000079",
    "uri": "s3://tv2v-data/raw/taxonomy/Task_198_Dashcam_Video_Super_Resolution/SpicyqSama007_nltk_data_bdd100k_3k/archives/val_00/bdd100k_videos_val_00.zip"
  },
  "media": {
    "input": {
      "file": "input_video.mp4",
      "width": 320,
      "height": 180,
      "fps": 30.0,
      "num_frames": 80
    },
    "target": {
      "file": "target_video.mp4",
      "width": 1280,
      "height": 720,
      "fps": 30.0,
      "num_frames": 80
    }
  },
  "pipeline": {
    "name": "dashcam_video_super_resolution",
    "version": "0.2.0",
    "commit": "c078a850f7f4c823d6057c7a6ec8a8298a87e51d"
  },
  "optional": {
    "source_member": "bdd100k/videos/val/b1c66a42-6f7d68ca.mov",
    "source_sha256": "a30203862234972202297b2b1fa65c7c8d69a20ea5ffe4b63f55f7c4053df825",
    "source_decoded_frames": 1196,
    "source_r_frame_rate": "60000/1001",
    "source_rotation_degrees": 90,
    "source_pixel_format": "yuv420p",
    "source_encoded_width": 720,
    "source_encoded_height": 1280,
    "window": {
      "ordinal": 0,
      "frame_start_inclusive": 0,
      "frame_end_exclusive": 80,
      "num_frames": 80,
      "target_frames": 81,
      "partition_policy": "nearest-window-count-balanced-full-coverage-v1"
    }
  },
  "identity": {
    "sample_fingerprint": "7a88f1592d052844efdd432d90ff85cae1ff332f74a1e16f7b150d4f471f6e54"
  }
}
Per-sample automated checks
CheckStatusSummary
task198.super-resolution-pairPASSInput is a faithful 320x180 area downsample of the aligned 1280x720 source-backed Target.

Required human checks

Exactly three checks are required for this sample. The active human may check only their own column.

QuestionOwnerHuman Evaluator
Did you completely watch this sample's Input and Target in the required playback mode?
Does the exact Prompt describe the intended task and the visible Input-to-Target relationship?
Are the metadata, automated result, and sample-specific viewing cues consistent with the media?
SAMPLE 07

Task_198_dashcam_video_super_resolution_000029195

test_0081-framesource-beginningrotation-90
not fully played

Source: BDD100K via SpicyqSama007/nltk_data · cabc30fc-e7726578/frames-000000-000080
Selection: Test-split exact-81 case checks held-out coverage.

Input and Target not completed.
INPUT · 81 frames · 320×180 · 30.0 fps
0:00 / --:--
TARGET · 81 frames · 1280×720 · 30.0 fps
0:00 / --:--
EXACT PROMPTRestore this low-resolution dashcam video to a sharp 1280x720 video while preserving the original scene, colors, motion, timing, and all visible content.
Watch for
  1. Watch the complete synchronized pair; confirm Input and Target have identical motion, framing, timing, and visible content.
  2. Confirm Target is genuinely sharper at 1280x720 while Input is the faithful 320x180 area downsample.
  3. Confirm smooth complete playback for this 81-frame near-81 window, including its first and last frame.
Complete exact metadata.json
{
  "schema_version": "1.2",
  "sample_id": "Task_198_dashcam_video_super_resolution_000029195",
  "task": {
    "id": "Task_198",
    "name": "dashcam_video_super_resolution",
    "modality": "v2v"
  },
  "source": {
    "dataset": "BDD100K via SpicyqSama007/nltk_data",
    "split": "test_00",
    "sample_id": "cabc30fc-e7726578/frames-000000-000080",
    "uri": "s3://tv2v-data/raw/taxonomy/Task_198_Dashcam_Video_Super_Resolution/SpicyqSama007_nltk_data_bdd100k_3k/archives/test_00/bdd100k_videos_test_00.zip"
  },
  "media": {
    "input": {
      "file": "input_video.mp4",
      "width": 320,
      "height": 180,
      "fps": 30.0,
      "num_frames": 81
    },
    "target": {
      "file": "target_video.mp4",
      "width": 1280,
      "height": 720,
      "fps": 30.0,
      "num_frames": 81
    }
  },
  "pipeline": {
    "name": "dashcam_video_super_resolution",
    "version": "0.2.0",
    "commit": "c078a850f7f4c823d6057c7a6ec8a8298a87e51d"
  },
  "optional": {
    "source_member": "bdd100k/videos/test/cabc30fc-e7726578.mov",
    "source_sha256": "300e42bb966f96a173488db2ca46e9f998bf1304da35d4bc5af281d613a148c4",
    "source_decoded_frames": 1203,
    "source_r_frame_rate": "30000/1001",
    "source_rotation_degrees": 90,
    "source_pixel_format": "yuv420p",
    "source_encoded_width": 720,
    "source_encoded_height": 1280,
    "window": {
      "ordinal": 0,
      "frame_start_inclusive": 0,
      "frame_end_exclusive": 81,
      "num_frames": 81,
      "target_frames": 81,
      "partition_policy": "nearest-window-count-balanced-full-coverage-v1"
    }
  },
  "identity": {
    "sample_fingerprint": "95f7d452de373c8d106a05e7ea2bb4b990e83ed2c38e9fb9354833ee19a17af5"
  }
}
Per-sample automated checks
CheckStatusSummary
task198.super-resolution-pairPASSInput is a faithful 320x180 area downsample of the aligned 1280x720 source-backed Target.

Required human checks

Exactly three checks are required for this sample. The active human may check only their own column.

QuestionOwnerHuman Evaluator
Did you completely watch this sample's Input and Target in the required playback mode?
Does the exact Prompt describe the intended task and the visible Input-to-Target relationship?
Are the metadata, automated result, and sample-specific viewing cues consistent with the media?
SAMPLE 08

Task_198_dashcam_video_super_resolution_000043805

test_0078-framesource-endingfinal-ordinalrotation-270
not fully played

Source: BDD100K via SpicyqSama007/nltk_data · cd35ea13-1ecb210c/frames-001105-001182
Selection: Final candidate proves deterministic census-tail and source-end coverage.

Input and Target not completed.
INPUT · 78 frames · 320×180 · 30.0 fps
0:00 / --:--
TARGET · 78 frames · 1280×720 · 30.0 fps
0:00 / --:--
EXACT PROMPTRestore this low-resolution dashcam video to a sharp 1280x720 video while preserving the original scene, colors, motion, timing, and all visible content.
Watch for
  1. Watch the complete synchronized pair; confirm Input and Target have identical motion, framing, timing, and visible content.
  2. Confirm Target is genuinely sharper at 1280x720 while Input is the faithful 320x180 area downsample.
  3. Confirm smooth complete playback for this 78-frame near-81 window, including its first and last frame.
Complete exact metadata.json
{
  "schema_version": "1.2",
  "sample_id": "Task_198_dashcam_video_super_resolution_000043805",
  "task": {
    "id": "Task_198",
    "name": "dashcam_video_super_resolution",
    "modality": "v2v"
  },
  "source": {
    "dataset": "BDD100K via SpicyqSama007/nltk_data",
    "split": "test_00",
    "sample_id": "cd35ea13-1ecb210c/frames-001105-001182",
    "uri": "s3://tv2v-data/raw/taxonomy/Task_198_Dashcam_Video_Super_Resolution/SpicyqSama007_nltk_data_bdd100k_3k/archives/test_00/bdd100k_videos_test_00.zip"
  },
  "media": {
    "input": {
      "file": "input_video.mp4",
      "width": 320,
      "height": 180,
      "fps": 30.0,
      "num_frames": 78
    },
    "target": {
      "file": "target_video.mp4",
      "width": 1280,
      "height": 720,
      "fps": 30.0,
      "num_frames": 78
    }
  },
  "pipeline": {
    "name": "dashcam_video_super_resolution",
    "version": "0.2.0",
    "commit": "c078a850f7f4c823d6057c7a6ec8a8298a87e51d"
  },
  "optional": {
    "source_member": "bdd100k/videos/test/cd35ea13-1ecb210c.mov",
    "source_sha256": "e5ff101f993f8357890d2b9da5fcf4c4d361e0ba09735b22d7839be0fafb6d76",
    "source_decoded_frames": 1183,
    "source_r_frame_rate": "30000/1001",
    "source_rotation_degrees": 270,
    "source_pixel_format": "yuv420p",
    "source_encoded_width": 720,
    "source_encoded_height": 1280,
    "window": {
      "ordinal": 14,
      "frame_start_inclusive": 1105,
      "frame_end_exclusive": 1183,
      "num_frames": 78,
      "target_frames": 81,
      "partition_policy": "nearest-window-count-balanced-full-coverage-v1"
    }
  },
  "identity": {
    "sample_fingerprint": "c1f151ffd151a1ee5522e869222c4c750539eff585e444a0a23627174f44c044"
  }
}
Per-sample automated checks
CheckStatusSummary
task198.super-resolution-pairPASSInput is a faithful 320x180 area downsample of the aligned 1280x720 source-backed Target.

Required human checks

Exactly three checks are required for this sample. The active human may check only their own column.

QuestionOwnerHuman Evaluator
Did you completely watch this sample's Input and Target in the required playback mode?
Does the exact Prompt describe the intended task and the visible Input-to-Target relationship?
Are the metadata, automated result, and sample-specific viewing cues consistent with the media?

One decision

Across all eight complete near-81 pairs, is the Target a visibly sharper, aligned 1280x720 version of the Input while preserving the exact scene, colours, motion, timing, and all visible content?

0 / 8 samples fully played; 0 / 24 required checks complete for the active role. PASS/FAIL remain disabled until all three checks are complete for every sample.

No decision selected.