talkingheadbench / tests /test_set /subenv1_cases.json
elix3r's picture
Upload folder using huggingface_hub
ab34aa7 verified
Raw
History Blame Contribute Delete
45.8 kB
{
"cases": [
{
"id": "001",
"source_file": "ref_001.PNG",
"observation": {
"face_occupancy_ratio": 0.26671083495020864,
"estimated_yaw_degrees": 0.20169640267008665,
"estimated_pitch_degrees": 7.253506203955002,
"background_complexity_score": 0.4984572520736105,
"lighting_uniformity_score": 0.24749941681606225,
"skin_tone_bucket": 3,
"occlusion_detected": false,
"image_resolution": [
1376,
768
],
"estimated_sharpness": 0.14986528113503222,
"prompt_token_count": 603,
"prompt_semantic_density": 0.7899159663865546,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "frontal_simple",
"acceptable_regimes": [],
"identified_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": 0.2017,
"computed_occupancy": 0.2667,
"computed_sharpness": 0.1499,
"suggested_regime": "frontal_simple",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.3598
}
}
},
{
"id": "002",
"source_file": "ref_002.PNG",
"observation": {
"face_occupancy_ratio": 0.19196479115635157,
"estimated_yaw_degrees": 1.1837981771982193,
"estimated_pitch_degrees": 8.53957457949557,
"background_complexity_score": 0.6743001595962432,
"lighting_uniformity_score": 0.3183147220868896,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1376,
768
],
"estimated_sharpness": 0.11192579470374425,
"prompt_token_count": 525,
"prompt_semantic_density": 0.8252427184466019,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "low_quality",
"acceptable_regimes": [
"frontal_simple"
],
"identified_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"crop reference image to show face occupying at least 40% of frame area before use",
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": 1.1838,
"computed_occupancy": 0.192,
"computed_sharpness": 0.1119,
"suggested_regime": "low_quality",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.3088
}
}
},
{
"id": "003",
"source_file": "ref_003.PNG",
"observation": {
"face_occupancy_ratio": 0.2725474081256173,
"estimated_yaw_degrees": 1.1255909122958667,
"estimated_pitch_degrees": 7.363127049464499,
"background_complexity_score": 0.832447068146124,
"lighting_uniformity_score": 0.06953988592873628,
"skin_tone_bucket": 3,
"occlusion_detected": false,
"image_resolution": [
1376,
768
],
"estimated_sharpness": 0.08162747156777021,
"prompt_token_count": 503,
"prompt_semantic_density": 0.7578947368421053,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "complex_background",
"acceptable_regimes": [
"frontal_simple"
],
"identified_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding",
"background complexity high β€” non-face regions compete for attention budget"
],
"valid_prompt_modifications": [
"re-shoot reference against a plain backdrop to prevent background tokens competing with face identity encoding",
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": 1.1256,
"computed_occupancy": 0.2725,
"computed_sharpness": 0.0816,
"suggested_regime": "complex_background",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.2455
}
}
},
{
"id": "004",
"source_file": "ref_004.PNG",
"observation": {
"face_occupancy_ratio": 0.25728518502278763,
"estimated_yaw_degrees": -0.7495898046909464,
"estimated_pitch_degrees": 8.439161133948497,
"background_complexity_score": 0.2291957932242163,
"lighting_uniformity_score": 0.2874764787508761,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1376,
768
],
"estimated_sharpness": 0.024452668892707457,
"prompt_token_count": 552,
"prompt_semantic_density": 0.7747747747747747,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "frontal_simple",
"acceptable_regimes": [],
"identified_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": -0.7496,
"computed_occupancy": 0.2573,
"computed_sharpness": 0.0245,
"suggested_regime": "frontal_simple",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.3937
}
}
},
{
"id": "005",
"source_file": "ref_005.PNG",
"observation": {
"face_occupancy_ratio": 0.22957304026931524,
"estimated_yaw_degrees": 0.2945575223997702,
"estimated_pitch_degrees": 7.296129818846514,
"background_complexity_score": 0.2749065557301927,
"lighting_uniformity_score": 0.2642211569809112,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1376,
768
],
"estimated_sharpness": 0.03503245213697283,
"prompt_token_count": 575,
"prompt_semantic_density": 0.8256880733944955,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "low_quality",
"acceptable_regimes": [
"frontal_simple"
],
"identified_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"crop reference image to show face occupying at least 40% of frame area before use",
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": 0.2946,
"computed_occupancy": 0.2296,
"computed_sharpness": 0.035,
"suggested_regime": "low_quality",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.3737
}
}
},
{
"id": "006",
"source_file": "ref_006.PNG",
"observation": {
"face_occupancy_ratio": 0.2955096691169522,
"estimated_yaw_degrees": 33.9946297521904,
"estimated_pitch_degrees": 5.352060047295191,
"background_complexity_score": 0.1681451286090439,
"lighting_uniformity_score": 0.33264004114925594,
"skin_tone_bucket": 5,
"occlusion_detected": false,
"image_resolution": [
1375,
768
],
"estimated_sharpness": 0.027245719034493597,
"prompt_token_count": 615,
"prompt_semantic_density": 0.819672131147541,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "non_frontal",
"acceptable_regimes": [],
"identified_risk_factors": [
"yaw exceeds 25Β° β€” lateral pose reduces reference token coverage",
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"use a frontal-facing reference image to maximise reference token coverage across all generation frames",
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": 33.9946,
"computed_occupancy": 0.2955,
"computed_sharpness": 0.0272,
"suggested_regime": "non_frontal",
"suggested_risk_factors": [
"yaw exceeds 25Β° β€” lateral pose reduces reference token coverage",
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.427
}
}
},
{
"id": "007",
"source_file": "ref_007.PNG",
"observation": {
"face_occupancy_ratio": 0.2709252152253281,
"estimated_yaw_degrees": 48.24818657014514,
"estimated_pitch_degrees": 6.081733989017453,
"background_complexity_score": 0.2410308726696404,
"lighting_uniformity_score": 0.3156295268313972,
"skin_tone_bucket": 5,
"occlusion_detected": false,
"image_resolution": [
1375,
768
],
"estimated_sharpness": 0.035017695319897346,
"prompt_token_count": 533,
"prompt_semantic_density": 0.8269230769230769,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "non_frontal",
"acceptable_regimes": [],
"identified_risk_factors": [
"yaw exceeds 25Β° β€” lateral pose reduces reference token coverage",
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"use a frontal-facing reference image to maximise reference token coverage across all generation frames",
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": 48.2482,
"computed_occupancy": 0.2709,
"computed_sharpness": 0.035,
"suggested_regime": "non_frontal",
"suggested_risk_factors": [
"yaw exceeds 25Β° β€” lateral pose reduces reference token coverage",
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.4032
}
}
},
{
"id": "008",
"source_file": "ref_008.PNG",
"observation": {
"face_occupancy_ratio": 0.3259303405880928,
"estimated_yaw_degrees": -71.20058423909946,
"estimated_pitch_degrees": 8.042769203847751,
"background_complexity_score": 0.40832051825647453,
"lighting_uniformity_score": 0.30459834592985213,
"skin_tone_bucket": 5,
"occlusion_detected": false,
"image_resolution": [
1375,
768
],
"estimated_sharpness": 0.05397594432379059,
"prompt_token_count": 623,
"prompt_semantic_density": 0.8048780487804879,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "non_frontal",
"acceptable_regimes": [],
"identified_risk_factors": [
"yaw exceeds 25Β° β€” lateral pose reduces reference token coverage",
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding",
"extreme lateral pose β€” profile view severely reduces usable reference tokens"
],
"valid_prompt_modifications": [
"use a frontal-facing reference image to maximise reference token coverage across all generation frames",
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": -71.2006,
"computed_occupancy": 0.3259,
"computed_sharpness": 0.054,
"suggested_regime": "non_frontal",
"suggested_risk_factors": [
"yaw exceeds 25Β° β€” lateral pose reduces reference token coverage",
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.3878
}
}
},
{
"id": "009",
"source_file": "ref_009.PNG",
"observation": {
"face_occupancy_ratio": 0.24534901604056358,
"estimated_yaw_degrees": -49.31794132251808,
"estimated_pitch_degrees": 8.126108627073185,
"background_complexity_score": 0.3477225107594732,
"lighting_uniformity_score": 0.2035173880281158,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1375,
768
],
"estimated_sharpness": 0.06152954527981137,
"prompt_token_count": 539,
"prompt_semantic_density": 0.822429906542056,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "low_quality",
"acceptable_regimes": [],
"identified_risk_factors": [
"yaw exceeds 25Β° β€” lateral pose reduces reference token coverage",
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"use a frontal-facing reference image to maximise reference token coverage across all generation frames",
"crop reference image to show face occupying at least 40% of frame area before use",
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": -49.3179,
"computed_occupancy": 0.2453,
"computed_sharpness": 0.0615,
"suggested_regime": "low_quality",
"suggested_risk_factors": [
"yaw exceeds 25Β° β€” lateral pose reduces reference token coverage",
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.3571
}
}
},
{
"id": "010",
"source_file": "ref_010.PNG",
"observation": {
"face_occupancy_ratio": 0.246126065030694,
"estimated_yaw_degrees": 0.7772138847854013,
"estimated_pitch_degrees": 1.4932629069629482,
"background_complexity_score": 1.0,
"lighting_uniformity_score": 0.34652263542053063,
"skin_tone_bucket": 3,
"occlusion_detected": false,
"image_resolution": [
1376,
768
],
"estimated_sharpness": 0.11093244484514717,
"prompt_token_count": 492,
"prompt_semantic_density": 0.7916666666666666,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "low_quality",
"acceptable_regimes": [
"frontal_simple"
],
"identified_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"crop reference image to show face occupying at least 40% of frame area before use",
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": 0.7772,
"computed_occupancy": 0.2461,
"computed_sharpness": 0.1109,
"suggested_regime": "low_quality",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.2653
}
}
},
{
"id": "011",
"source_file": "ref_011.PNG",
"observation": {
"face_occupancy_ratio": 0.27394223301248116,
"estimated_yaw_degrees": 0.21367615394448689,
"estimated_pitch_degrees": 9.404006680656586,
"background_complexity_score": 0.1432163076682715,
"lighting_uniformity_score": 0.3166478441512127,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1375,
768
],
"estimated_sharpness": 0.027717897432831644,
"prompt_token_count": 530,
"prompt_semantic_density": 0.7757009345794392,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "frontal_simple",
"acceptable_regimes": [],
"identified_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": 0.2137,
"computed_occupancy": 0.2739,
"computed_sharpness": 0.0277,
"suggested_regime": "frontal_simple",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.4224
}
}
},
{
"id": "012",
"source_file": "ref_012.PNG",
"observation": {
"face_occupancy_ratio": 0.9752188069479806,
"estimated_yaw_degrees": 1.674451900261819,
"estimated_pitch_degrees": 6.958791072623426,
"background_complexity_score": 0.0,
"lighting_uniformity_score": 0.14257801088947297,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1375,
768
],
"estimated_sharpness": 0.06960971678870148,
"prompt_token_count": 492,
"prompt_semantic_density": 0.8181818181818182,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "frontal_simple",
"acceptable_regimes": [],
"identified_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": 1.6745,
"computed_occupancy": 0.9752,
"computed_sharpness": 0.0696,
"suggested_regime": "frontal_simple",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.635
}
}
},
{
"id": "013",
"source_file": "ref_013.PNG",
"observation": {
"face_occupancy_ratio": 0.35466248867186634,
"estimated_yaw_degrees": -26.30671239208516,
"estimated_pitch_degrees": 11.114116703846703,
"background_complexity_score": 0.0983039892281269,
"lighting_uniformity_score": 0.3305038067695233,
"skin_tone_bucket": 5,
"occlusion_detected": false,
"image_resolution": [
1376,
768
],
"estimated_sharpness": 0.012672271111359932,
"prompt_token_count": 508,
"prompt_semantic_density": 0.8584905660377359,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "non_frontal",
"acceptable_regimes": [
"frontal_simple"
],
"identified_risk_factors": [
"yaw exceeds 25Β° β€” lateral pose reduces reference token coverage",
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"use a frontal-facing reference image to maximise reference token coverage across all generation frames",
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": -26.3067,
"computed_occupancy": 0.3547,
"computed_sharpness": 0.0127,
"suggested_regime": "non_frontal",
"suggested_risk_factors": [
"yaw exceeds 25Β° β€” lateral pose reduces reference token coverage",
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.4554
}
}
},
{
"id": "014",
"source_file": "ref_014.PNG",
"observation": {
"face_occupancy_ratio": 0.2697721481323242,
"estimated_yaw_degrees": -0.3540482436055811,
"estimated_pitch_degrees": 9.357075790533312,
"background_complexity_score": 0.11475734441122168,
"lighting_uniformity_score": 0.2200751983330147,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1375,
768
],
"estimated_sharpness": 0.029296646734595873,
"prompt_token_count": 568,
"prompt_semantic_density": 0.8141592920353983,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "frontal_simple",
"acceptable_regimes": [],
"identified_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": -0.354,
"computed_occupancy": 0.2698,
"computed_sharpness": 0.0293,
"suggested_regime": "frontal_simple",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.4079
}
}
},
{
"id": "015",
"source_file": "ref_015.PNG",
"observation": {
"face_occupancy_ratio": 0.30095508681102234,
"estimated_yaw_degrees": 0.0787174722887595,
"estimated_pitch_degrees": 6.783640210215074,
"background_complexity_score": 0.14719240547615264,
"lighting_uniformity_score": 0.20841154811291385,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1375,
768
],
"estimated_sharpness": 0.03332163711071112,
"prompt_token_count": 611,
"prompt_semantic_density": 0.7751937984496124,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "frontal_simple",
"acceptable_regimes": [],
"identified_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": 0.0787,
"computed_occupancy": 0.301,
"computed_sharpness": 0.0333,
"suggested_regime": "frontal_simple",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.4092
}
}
},
{
"id": "016",
"source_file": "ref_016.PNG",
"observation": {
"face_occupancy_ratio": 0.31093581007285553,
"estimated_yaw_degrees": -1.0653470539799819,
"estimated_pitch_degrees": 9.178804381827895,
"background_complexity_score": 0.09495288573553234,
"lighting_uniformity_score": 0.2792656267313851,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1375,
768
],
"estimated_sharpness": 0.014105671392825325,
"prompt_token_count": 537,
"prompt_semantic_density": 0.7962962962962963,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "frontal_simple",
"acceptable_regimes": [],
"identified_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": -1.0653,
"computed_occupancy": 0.3109,
"computed_sharpness": 0.0141,
"suggested_regime": "frontal_simple",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.433
}
}
},
{
"id": "017",
"source_file": "ref_017.PNG",
"observation": {
"face_occupancy_ratio": 0.9529492173876082,
"estimated_yaw_degrees": 0.37582876566065415,
"estimated_pitch_degrees": 3.96832173035298,
"background_complexity_score": 0.024831383373851972,
"lighting_uniformity_score": 0.24935375608227717,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1374,
768
],
"estimated_sharpness": 0.048195612447065564,
"prompt_token_count": 540,
"prompt_semantic_density": 0.7413793103448276,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "frontal_simple",
"acceptable_regimes": [],
"identified_risk_factors": [],
"valid_prompt_modifications": [
"add explicit identity descriptor such as 'portrait of OHWXPERSON' to strengthen prompt-to-reference binding"
],
"_annotation_notes": {
"computed_yaw": 0.3758,
"computed_occupancy": 0.9529,
"computed_sharpness": 0.0482,
"suggested_regime": "frontal_simple",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.6404
}
}
},
{
"id": "018",
"source_file": "ref_018.PNG",
"observation": {
"face_occupancy_ratio": 0.9529492173876082,
"estimated_yaw_degrees": 0.37582876566065415,
"estimated_pitch_degrees": 3.96832173035298,
"background_complexity_score": 0.024831383373851972,
"lighting_uniformity_score": 0.24935375608227717,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1374,
768
],
"estimated_sharpness": 0.048195612447065564,
"prompt_token_count": 559,
"prompt_semantic_density": 0.7631578947368421,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "frontal_simple",
"acceptable_regimes": [],
"identified_risk_factors": [],
"valid_prompt_modifications": [
"add explicit identity descriptor such as 'portrait of OHWXPERSON' to strengthen prompt-to-reference binding"
],
"_annotation_notes": {
"computed_yaw": 0.3758,
"computed_occupancy": 0.9529,
"computed_sharpness": 0.0482,
"suggested_regime": "frontal_simple",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.6404
}
}
},
{
"id": "019",
"source_file": "ref_019.PNG",
"observation": {
"face_occupancy_ratio": 0.9529492173876082,
"estimated_yaw_degrees": 0.37582876566065415,
"estimated_pitch_degrees": 3.96832173035298,
"background_complexity_score": 0.024831383373851972,
"lighting_uniformity_score": 0.24935375608227717,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1374,
768
],
"estimated_sharpness": 0.048195612447065564,
"prompt_token_count": 628,
"prompt_semantic_density": 0.7906976744186046,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "frontal_simple",
"acceptable_regimes": [],
"identified_risk_factors": [],
"valid_prompt_modifications": [
"add explicit identity descriptor such as 'portrait of OHWXPERSON' to strengthen prompt-to-reference binding"
],
"_annotation_notes": {
"computed_yaw": 0.3758,
"computed_occupancy": 0.9529,
"computed_sharpness": 0.0482,
"suggested_regime": "frontal_simple",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.6404
}
}
},
{
"id": "020",
"source_file": "ref_020.PNG",
"observation": {
"face_occupancy_ratio": 0.9529492173876082,
"estimated_yaw_degrees": 0.37582876566065415,
"estimated_pitch_degrees": 3.96832173035298,
"background_complexity_score": 0.024831383373851972,
"lighting_uniformity_score": 0.24935375608227717,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1374,
768
],
"estimated_sharpness": 0.048195612447065564,
"prompt_token_count": 577,
"prompt_semantic_density": 0.7394957983193278,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "frontal_simple",
"acceptable_regimes": [],
"identified_risk_factors": [],
"valid_prompt_modifications": [
"add explicit identity descriptor such as 'portrait of OHWXPERSON' to strengthen prompt-to-reference binding"
],
"_annotation_notes": {
"computed_yaw": 0.3758,
"computed_occupancy": 0.9529,
"computed_sharpness": 0.0482,
"suggested_regime": "frontal_simple",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.6404
}
}
},
{
"id": "021",
"source_file": "ref_021.PNG",
"observation": {
"face_occupancy_ratio": 0.304949886622754,
"estimated_yaw_degrees": -0.15055316337723074,
"estimated_pitch_degrees": 2.542070855174985,
"background_complexity_score": 1.0,
"lighting_uniformity_score": 0.11235041200594509,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1376,
768
],
"estimated_sharpness": 0.15576428583752805,
"prompt_token_count": 609,
"prompt_semantic_density": 0.8032786885245902,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "complex_background",
"acceptable_regimes": [
"frontal_simple"
],
"identified_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding",
"background complexity high β€” non-face regions compete for attention budget"
],
"valid_prompt_modifications": [
"re-shoot reference against a plain backdrop to prevent background tokens competing with face identity encoding",
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": -0.1506,
"computed_occupancy": 0.3049,
"computed_sharpness": 0.1558,
"suggested_regime": "complex_background",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.2451
}
}
},
{
"id": "022",
"source_file": "ref_022.PNG",
"observation": {
"face_occupancy_ratio": 0.22946405224502087,
"estimated_yaw_degrees": -34.15498902609289,
"estimated_pitch_degrees": 7.65763616375718,
"background_complexity_score": 0.11433578689099194,
"lighting_uniformity_score": 0.23727483235973457,
"skin_tone_bucket": 4,
"occlusion_detected": false,
"image_resolution": [
1375,
768
],
"estimated_sharpness": 0.026941647266423074,
"prompt_token_count": 631,
"prompt_semantic_density": 0.8114754098360656,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "low_quality",
"acceptable_regimes": [],
"identified_risk_factors": [
"yaw exceeds 25Β° β€” lateral pose reduces reference token coverage",
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"use a frontal-facing reference image to maximise reference token coverage across all generation frames",
"crop reference image to show face occupying at least 40% of frame area before use",
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": -34.155,
"computed_occupancy": 0.2295,
"computed_sharpness": 0.0269,
"suggested_regime": "low_quality",
"suggested_risk_factors": [
"yaw exceeds 25Β° β€” lateral pose reduces reference token coverage",
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.3988
}
}
},
{
"id": "023",
"source_file": "ref_023.PNG",
"observation": {
"face_occupancy_ratio": 0.28388736112551255,
"estimated_yaw_degrees": -0.7162607505192292,
"estimated_pitch_degrees": 6.532766043811455,
"background_complexity_score": 1.0,
"lighting_uniformity_score": 0.18511442845223403,
"skin_tone_bucket": 3,
"occlusion_detected": false,
"image_resolution": [
1376,
768
],
"estimated_sharpness": 0.21969819608933425,
"prompt_token_count": 580,
"prompt_semantic_density": 0.8260869565217391,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "complex_background",
"acceptable_regimes": [
"frontal_simple"
],
"identified_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding",
"background complexity high β€” non-face regions compete for attention budget"
],
"valid_prompt_modifications": [
"re-shoot reference against a plain backdrop to prevent background tokens competing with face identity encoding",
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": -0.7163,
"computed_occupancy": 0.2839,
"computed_sharpness": 0.2197,
"suggested_regime": "complex_background",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.2661
}
}
},
{
"id": "024",
"source_file": "ref_024.PNG",
"observation": {
"face_occupancy_ratio": 0.27887891151688315,
"estimated_yaw_degrees": -16.23611516297194,
"estimated_pitch_degrees": 5.286140894024532,
"background_complexity_score": 0.09903928152648972,
"lighting_uniformity_score": 0.3084112080049588,
"skin_tone_bucket": 5,
"occlusion_detected": false,
"image_resolution": [
1376,
768
],
"estimated_sharpness": 0.01690383733356814,
"prompt_token_count": 466,
"prompt_semantic_density": 0.7525773195876289,
"conflicting_descriptors": [],
"identity_anchoring_strength": 0.5
},
"ground_truth": {
"regime_classification": "frontal_simple",
"acceptable_regimes": [
"non_frontal"
],
"identified_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"valid_prompt_modifications": [
"ensure even lighting on face β€” avoid strong directional light that creates shadow regions"
],
"_annotation_notes": {
"computed_yaw": -16.2361,
"computed_occupancy": 0.2789,
"computed_sharpness": 0.0169,
"suggested_regime": "frontal_simple",
"suggested_risk_factors": [
"lighting uniformity low β€” shadow regions risk inconsistent skin rendering",
"face occupancy below 0.4 β€” background competes for attention budget",
"low sharpness β€” may cause blurred identity encoding"
],
"image_usability_score": 0.4289
}
}
}
]
}