Image Library Analytics
5107 images · analyzed entirely on-device · SigLIP 2 · aesthetic-predictor-v2-5 · CLIP-IQA+ · DINOv2 · RMBG-2.0
This report analyzes your photo library using on-device AI models — covering shooting habits, composition patterns, editing style, color grading, and technical quality.
Highlights
Dominant Scene
people and portraits
45% of library
Avg Aesthetic Score
62.8 / 100
range 31–81
Visual Clusters
10
DINOv2 + KMeans
Avg Resolution
24.1MP
2.67–48.77MP range
Exposure Bias
-0.08
0 = neutral
Editing Style
Varied / Eclectic
91% confidence
Redundant Images
169
similar image duplicates
ELA Outliers
1
unusual compression artifacts
ISO Preference
125
mixed-light shooter
Prime vs Zoom
98% prime
2% zoom lenses
Capture Style
Fast Shutter
64% of shots
Avg IQ Score
58/100
sharpness · exposure · noise
High Feel, Low IQ
51 photos
aesthetic >65 · technical IQ <40
Issues Summary
Edge density in the bottom quarter of the frame is high, suggesting a cluttered or distracting foreground. Move closer, change angle, or use a wider aperture to blur foreground elements.
Many shots have a tilted horizon or roofline — even a 1–2° tilt reads as unsteady. Enable your camera's electronic level while shooting, or use the Crop & Rotate tool in Lightroom to straighten affected images.
The center of the frame is no sharper than the surroundings — the subject blends into the background. Try a wider aperture (lower f-number) for shallower depth of field, or reframe to create tonal contrast between subject and background.
The main subject falls near the center of the frame rather than on a rule-of-thirds intersection. Try placing your subject at one of the four power points (roughly 1/3 from any edge) to add visual tension and lead the eye.
51 photos score high aesthetically (>65) but low on technical quality (<40) — likely due to motion blur, noise, or soft focus. These are candidates for a reshoot or a sharper edit pass.
What You Shoot
Scene Preferences
45%
27%
14%
4%
4%
3%
2%
1%
Visual Attributes & Aesthetic Florence-2 VQA · % share · avg aesthetic score
Season
autumn
52.3% · 63.2
summer
22.3% · 61.2
spring
12.7% · 60.9
winter
12.6% · 65.7
People Present
no
53.3% · 62.7
yes
46.7% · 62.8
Setting
outdoors
81.2% · 64.1
indoors
18.8% · 57.0
Time of Day
afternoon
36.0% · 64.5
night
25.4% · 60.3
evening
22.9% · 62.7
morning
15.7% · 62.8
Weather
rainy
44.2% · 61.2
clear
42.0% · 63.5
snowy
7.3% · 65.6
foggy
3.7% · 65.3
cloudy
2.8% · 64.5
Visual Similarity Map 800 of 5107 images · hover for details
Each dot is one photo. Nearby dots look visually similar. Clusters reveal shooting tendencies — tight groups mean a consistent style, isolated dots are unique shots. Colored by scene type; hover for details.
How You Shoot
Focal Length
Aperture
ISO Sensitivity
Composition Style
88.0%
Rule of Thirds
Score: 0.83 — 0=centred, 1=off-centre
Subject or key element placed near a thirds intersection, adding visual tension and naturally directing the viewer's eye across the frame.
42.2%
Subject Isolation
Score: 1.49 — ratio vs surroundings
Centre zone is significantly sharper than the surroundings — the subject pops cleanly from its background via shallow depth of field or high contrast.
41.9%
Strong Symmetry
Score: 0.85 — left/right mirror ratio
High left/right mirroring across the frame, common in architectural subjects, reflections, or deliberately geometric compositions.
36.0%
Negative Space
Score: 0.86 — empty-frame fraction
Unusually large empty areas surround the subject, giving it room to breathe and drawing the viewer's attention through contrast with the void.
30.4%
Foreground Interest
Score: 0.15 — lower = cleaner frame
Complex edge density in the bottom zone indicates foreground elements that add depth, texture, or layering to the scene.
Composition Grid
Detail Density by Zone
0.06
0.07
0.06
0.12
0.17
0.13
0.12
0.14
0.12
Brighter = more edges and detail in that zone. Centre 0.17 vs corners avg 0.09.
Subject Placement
0%
1%
0%
2%
62%
3%
3%
26%
3%
Darker green = more images placed in that zone. Subject fills avg 41% of the frame; 0.145 off-centre (0 = perfectly centred).
Horizon Position ?
Zone Interpretation
Center-dominant — most visual detail and subject matter sits in the middle of the frame. The centre zone scores 0.17 versus a corner average of 0.09, a ratio typical of telephoto or portrait-heavy libraries.
Sharpness & Exposure
Very Sharp
Avg Sharpness
score 1087 · >800 = crisp, <150 = blurry
-0.08
Exposure Bias
0 = neutral · + = brighter · − = darker
8.1%
Highlights Clipped
images with blown highlights
Portrait Insights 2385 images · YOLO11n-pose
Solo
1556
1 person
Pair
368
2 people
Group 3+
461
3 or more
Framing Style
65%
tight
63.4 aesthetic
26%
medium
61.6 aesthetic
9%
wide
60.0 aesthetic
Gear & Devices
Camera Bodies & Lenses
Camera body types across the library.
4939
mirrorless
62.8 aes
167
phone
61.8 aes
1
unknown
66.3 aes
Top Lenses
2396
2214
137
102
96
60
Motion Style
Fast Shutter (≤1/500s)
63.7%
Handheld (1/500–1/60s)
34.6%
Slow (1/60–1/15s)
1.0%
Bulb (>1/15s)
0.7%
Editing & Aesthetics
Editing Style Analysis
Varied / Eclectic
91% confidence
Your images don't follow a single visual recipe — saturation, contrast, and brightness vary widely. This usually means different shoots, different subjects, and multiple creative moods rather than one applied preset. It's a sign of range, not inconsistency.
Aesthetic Quality aesthetic-predictor-v2-5
Average
62.8
31–81 range · out of 100
Median
63.5
Std Dev
7.00
lower = more uniform
Scores of 50–65 indicate well-edited photography. Your avg of 62.8 places your library solidly in that bracket.
Color Grading
Tonal Style
flat_matte
Compressed tonal range — lifted shadows, rolled highlights; the 'matte look'.
Color Temp.
neutral
No strong colour temperature bias (~5500K). Daylight-balanced or mixed conditions.
Color Harmony
Split Toning Fingerprint
Shadows
#343e32
Midtones
#8c8372
Highlights
#c5cdc0
Green highlights · green shadows
Balance -1.6 · shadow lift 19.1 EV · highlight push 4.8 EV
Color & Mood
Subject
Background
Saturation
0.23
eclectic
Brightness
0.51
0 dark · 1 bright
Hue Distribution
dominant color ranges across your library (chromatic images, sat > 15%)
Saturation Distribution
48%
42%
8%
1%
MutedVivid
Signature Edit
Heavy tonal recovery — pulling highlights (-72) while lifting shadows (+60) for maximum dynamic range. Rich selective colour — vibrance +40 over global saturation (-15) keeps skin tones natural. Brightening exposure by 0.5 stop.
Tone
Exposure+0.5 ±1
Highlights-72.0 ±35
Shadows+60.0 ±30
Whites-58.0 ±35
Blacks+3.0 ±13
Contrast+5.0 ±16
Presence & Color
Texture-5.0 ±10
Dehaze+0.0 ±9
Vibrance+40.0 ±22
Saturation-15.0 ±13
White Balance6105K · warm
Editing DNA HSL channel fingerprint · avg across Lightroom library
How you systematically shift individual colour channels in post — your tonal and colour signature.
Hue
Sat
Lum
Red
-0.2
-6.0
+0.9
Orange
-5.8
-6.8
+11.2
Yellow
-27.8
-21.5
+1.6
Green
-29.5
-28.9
+1.7
Aqua
-6.3
-24.2
+0.6
Blue
-16.5
-28.1
+3.6
Purple
-9.5
-26.7
+0.2
Magenta
-1.2
-16.5
+0.4
Purple = raised · Red = lowered · Dark = near-zero
Adobe Lightroom Develop Settings 2917 of 5107 images with develop metadata
Avg slider positions across your library. Only images synced from Lightroom include develop settings — local folder imports don't carry this data. Lum. Noise = Luminance Smoothing, which reduces grain/texture noise in the brightness channel (high-ISO shots). Color NR = Color Noise Reduction, which removes coloured speckles (chroma noise) in shadow areas.
Contrast
+25.0
Shadows
+5.0
Vibrance
+30.3
Saturation
-14.5
Sharpness
+43.4
Lum. Noise
+5.4
Color NR
+23.5
Avg Slider Value by Scene ▾
| +25 | +25 | — | +25 | +25 | +25 | — | +25 | — | — |
| +29 | +34 | +29 | +18 | +19 | +36 | +32 | +26 | +20 | +27 |
| -14 | -15 | -12 | -8 | -6 | -20 | -15 | -16 | -6 | -13 |
| +45 | +41 | +52 | +51 | +43 | +47 | +41 | +47 | +46 | +36 |
| +5 | +3 | +9 | +4 | +6 | +4 | +3 | +26 | +3 | +2 |
| +24 | +22 | +27 | +22 | +25 | +25 | +25 | +28 | +20 | +23 |
Green = pushed above default · Red = pulled below · Intensity = magnitude of avg shift
Color Profile, Aesthetic Score & Editing Intensity by Scene
nature
Warm
⬡ 65.3
IQ 59.6
σ 6.32
✏ 240
sat 0.23
bri 0.54
1389 images
travel and landmarks
Warm
⬡ 64.6
IQ 67.8
σ 4.81
✏ 317
sat 0.26
bri 0.71
13 images
architecture
Cool
⬡ 62.8
IQ 66.6
σ 6.33
✏ 201
sat 0.18
bri 0.58
739 images
animals
Warm
⬡ 62.2
IQ 60.7
σ 5.23
✏ 263
sat 0.23
bri 0.44
210 images
people and portraits
Warm
⬡ 62.0
IQ 56.2
σ 7.04
✏ 255
sat 0.24
bri 0.51
2274 images
urban street
Warm
⬡ 61.5
IQ 61.4
σ 5.79
✏ 223
sat 0.18
bri 0.52
139 images
night scene
Cool
⬡ 59.8
IQ 46.7
σ 7.55
✏ 186
sat 0.29
bri 0.20
86 images
abstract
Warm
⬡ 59.0
IQ 45.7
σ 9.05
✏ 244
sat 0.27
bri 0.29
212 images
food
Warm
⬡ 57.9
IQ 51.9
σ 3.65
✏ 168
sat 0.31
bri 0.37
10 images
interior
Warm
⬡ 53.0
IQ 55.1
σ 6.84
✏ 164
sat 0.17
bri 0.46
35 images
⬡ = avg aesthetic score · IQ = avg technical quality · σ = consistency (lower = more uniform) · ✏ = editing intensity
Editing Trends Over Time
early
63.3
avg aesthetic
161 intensity
5% automated
recent
61.3
avg aesthetic
164 intensity
0% automated
Portfolio Albums Adobe Lightroom collections · curated selections
Images you've placed in albums represent your curated portfolio — the subset you've actively chosen to showcase. Comparing album images against the rest of your library shows how well the AI aesthetic model aligns with your own taste.
16
Albums
Lightroom collections
738
Curated Images
of 5107 Lightroom images
14.5%
Curation Rate
in at least one album
AI Aesthetic Agreement
63.5
avg score · in albums
vs
62.6
avg score · not in albums
+0.8
model bias
The model leans toward agreeing with your selections — a modest positive bias.
Show album details ▾
| Album | Images | Avg Aesthetic | σ | Edit Intensity | Palette |
|---|---|---|---|---|---|
| Nyc | 88 | 65.2 | 3.90 | — | |
| Yellowstone | 83 | 67.7 | 4.73 | — | |
| Bayarea | 75 | 66.0 | 5.39 | 300 | |
| Flora & Fauna | 74 | 62.8 | 4.17 | 280 | |
| Misc | 71 | 62.1 | 5.83 | 198 | |
| Dubai | 57 | 61.8 | 5.52 | 329 | |
| Varanasi | 55 | 62.1 | 3.78 | — | |
| Art | 50 | 55.3 | 4.68 | 273 | |
| Iphone | 48 | 59.4 | 8.41 | 75 | |
| New England | 39 | 69.4 | 4.83 | 310 | |
| Utah & Vegas | 36 | 65.3 | 5.52 | 326 | |
| Mexico | 27 | 63.1 | 5.38 | 343 | |
| Texas | 18 | 63.2 | 4.95 | 301 | |
| Mom | 12 | 55.4 | 2.78 | — | |
| Arizona | 12 | 68.2 | 1.92 | 328 | |
| Oregon | 9 | 69.5 | 5.45 | 388 |
Editing Journey
Editing Journey Over Time
Workflow Adoption % of images per month using each technique · includes automation score
Edit Recency how soon after capture were photos edited
3%
Same Day
131 images
25%
Within Week
1264 images
43%
Within Month
2209 images
29%
Later
1503 images
When You Shoot
Shooting Hours golden hour = 5–7 AM/PM highlighted
Monthly Shooting
Day of Week Lightroom capture dates
Shooting Profile
Shooting Profile by Context time of day × scene — typical gear and results
| Time | Scene | N | Focal (mm) | f/ | ISO | Avg Aes | Best Combo |
|---|---|---|---|---|---|---|---|
| night | abstract | 7 | 85 | 2.8 | 125 | 60.1 | telephoto+shallow |
| night | animals | 12 | 85 | 1.8 | 100 | 63.1 | telephoto+shallow |
| night | architecture | 147 | 50 | 2.8 | 125 | 60.5 | normal+mid |
| night | interior | 3 | 50 | 1.8 | 200 | 55.9 | normal+shallow |
| night | nature | 75 | 85 | 1.8 | 100 | 67.2 | telephoto+shallow |
Photo Events time gaps > 4 h = new event
Shooting sessions grouped by capture time — events separated by gaps >4 hours.
120
Events
distinct shooting sessions
42.3
Avg Images
per event
363
Largest Event
images in biggest session
| Date | Scene | Images | Duration |
|---|---|---|---|
| 2025-03-02 | architecture | 267 | 4.9h |
| 2025-01-27 | people and portraits | 111 | 3.6h |
| 2025-01-26 | people and portraits | 121 | 1.6h |
| 2024-09-01 | architecture | 67 | 4.3h |
| 2024-08-31 | architecture | 120 | 9.8h |
| 2024-07-28 | people and portraits | 76 | 4.8h |
| 2024-07-05 | nature | 120 | 9.3h |
| 2024-07-04 | nature | 202 | 12.1h |
| 2024-07-03 | nature | 57 | 1.1h |
| 2024-05-26 | architecture | 94 | 7.2h |
| 2024-05-25 | people and portraits | 131 | 4.5h |
Similar Image Groups DINOv2 · DBSCAN ε=0.05
Images grouped by visual similarity using DINOv2 embeddings and DBSCAN clustering — the best-scoring image in each group is the keep candidate; the rest are safe to cull. Top 10 largest groups shown.
120
Groups Found
similarity clusters
169
Redundant Images
safe to cull
3.3%
Redundancy Rate
of analysed library
4818
Unique Images
no similar match found
Burst Shooting
Bursts
120
shooting sessions
In Bursts
289
5.7% of library
Largest
17
shots in one burst
Forensics
ELA Forensics JPEG compression analysis
Error Level Analysis flags regions with inconsistent JPEG compression — a relative signal for local edits or compositing. Lightroom-exported JPEGs naturally score higher than raw camera originals; use the extreme outlier count as a within-library signal, not an absolute threshold.
5107
JPEGs Analyzed
non-JPEG files skipped
60.98
Avg Max Error
library baseline
1
Extreme Outliers
max error > 171.0 (3×IQR above Q3)
255
ELA/IQ Conflict
high compression artifact · low IQ score
Quality Issues 5107 images · local rule engine
30.4%
Busy / cluttered foreground
1551 images · nature
Edge density in the bottom quarter of the frame is high, suggesting a cluttered or distracting foreground. Move closer, change angle, or use a wider aperture to blur foreground elements.
18.2%
Horizon not level
929 images · people and portraits
Many shots have a tilted horizon or roofline — even a 1–2° tilt reads as unsteady. Enable your camera's electronic level while shooting, or use the Crop & Rotate tool in Lightroom to straighten affected images.
13.1%
Subject not well isolated from background
669 images · nature
The center of the frame is no sharper than the surroundings — the subject blends into the background. Try a wider aperture (lower f-number) for shallower depth of field, or reframe to create tonal contrast between subject and background.
9.6%
Subject too centered
492 images · people and portraits
The main subject falls near the center of the frame rather than on a rule-of-thirds intersection. Try placing your subject at one of the four power points (roughly 1/3 from any edge) to add visual tension and lead the eye.
6.3%
Highlight clipping — overexposed areas
324 images · people and portraits
More than 3% of pixels are blown to pure white (brightness > 250). Highlight clipping is irreversible — detail in those areas is lost. Dial in −1 to −2 stops of exposure compensation in bright scenes, or use spot metering on the highlights.
2.3%
Shadow clipping — crushed blacks
119 images · people and portraits
More than 5% of pixels are at pure black (brightness < 5), losing shadow detail. This can be intentional for high-contrast moody shots, but if unintended, lift the shadows or use fill flash. Shooting RAW retains more recoverable shadow data.
1.3%
Portrait with deep depth-of-field
65 images · people and portraits
EXIF data shows a small aperture (f/8+) on a portrait, meaning the background is in sharp focus and competes with the subject. For environmental portraits this can work, but for headshots/closeups, f/1.4–f/2.8 will separate the subject more cleanly.
1.0%
Strong feel, weak technical quality
51 images · people and portraits
CLIP-IQA+ rates this photo technically poor (noise, blur, compression) but the aesthetic score is high — the composition and mood are strong. This is a reshoot candidate: the same scene with better technique could produce a standout image.
0.7%
Subject off-center but not on a thirds line
38 images · people and portraits
The saliency map places the subject noticeably off-center, but the thirds score suggests it's not landing on a power point either. Try reframing to a true rule-of-thirds position, or anchor the subject at centre for a deliberate symmetric composition.
0.1%
Low overall aesthetic score
4 images · people and portraits
The aesthetic-predictor-v2-5 (trained on human aesthetic ratings) scored this photo below 35/100. Common causes: flat or cluttered composition, poor exposure, heavy noise, or an uninteresting subject. Typical edited photos score 40–60; professional work 60+.
0.0%
Low-light shot with low aesthetic score
1 images · people and portraits
Shot at ISO > 1600 and the aesthetic score is below 40 — likely affected by noise, motion blur, or flat exposure. Consider a faster lens, stabilisation, or expose-to-the-right and denoise in post.