Unit 6: Advancement in Impression Analysis

I. Orientation: Impression Evidence and the Digital Turn

Impression evidence is any two- or three-dimensional mark left when one object presses against a receiving surface (skin ridges, lips, teeth, footwear, tyres, tool edges). Modern analysis has shifted from purely manual comparison toward digital capture, computational enhancement and searchable databases, dramatically improving reproducibility and evidentiary weight.

• Class vs. individual characteristics: class features (pattern type, size, mould) narrow the source population; individual features (accidental damage, minutiae, pore positions) allow identification to a single source.
• Positive vs. negative impressions: positive impressions transfer a substance (blood, dust); negative impressions are indentations in a plastic medium (soil, putty).
• Latent, patent, plastic: latent marks are invisible until developed; patent are visible; plastic are three-dimensional.
• ACE-V framework: Analysis, Comparison, Evaluation, Verification — the reasoning chain all impression disciplines share.
• Reproducibility mandate: every enhancement step must be logged and reversible so the original data survives court scrutiny (Daubert-style admissibility).

II. Utilization of Forensic Photography in Impression Analysis

Photography converts a fragile physical mark into a permanent, scaled, admissible record before any invasive processing.

A. Purpose and Principle

• Documentation before development: capture the mark as-found, then again after chemical or physical enhancement, preserving chain-of-custody imagery.
• 1:1 scaling: a reference scale (ABFO No. 2 ruler) is placed coplanar with the impression so measurements are recoverable from the image.

B. Utilization of Forensic Photography in Impression Analysis

• Oblique (raking) light: a low-angle beam (roughly 5°–30°) casts shadows across ridge or shoeprint relief, boosting contrast on dust and indented marks.
• Ultraviolet and infrared bands: UV (≈254–365 nm) reveals residues that fluoresce; IR penetrates interfering backgrounds like dark or patterned surfaces.
• Alternate light sources (ALS): tunable wavelengths (450 nm blue, 530 nm green) with barrier filters excite treated latent prints (e.g., cyanoacrylate stained with basic yellow).
• Perpendicular alignment: the film/sensor plane is kept parallel to the impression plane to eliminate perspective distortion, essential for footwear length metrics.
• Fixed camera platforms: copy stands and tripods with cable release avoid motion blur during long UV/IR exposures.

III. Image Enhancement

Image enhancement mathematically improves the visibility of ridge or pattern detail without fabricating features that were not present.

A. Definition and Constraint

• Non-additive rule: enhancement may suppress noise or boost existing signal but must never introduce data; the workflow file records each operation.

B. Image Enhancement

• Histogram adjustment: stretching the intensity histogram redistributes pixel values so a faint mark spans the full 0–255 range.
• Fourier transform filtering: converting to the frequency domain isolates a repetitive background (fabric weave, banknote guilloché) as discrete peaks that are masked, then inverse-transformed to leave the aperiodic print.

enhanced = IFFT( FFT(image) × filter_mask )

• FFT: Fast Fourier Transform, mapping spatial data to frequency space.
• filter_mask: band-reject array set to 0 at background frequencies, 1 elsewhere.
  • Colour channel separation: splitting into R, G, B (or HSV) selects the channel where a bloody print contrasts most against its substrate.
  • Contrast-limited adaptive histogram equalisation (CLAHE): enhances local contrast in tiles while capping amplification to avoid noise blow-up.
  • Convolution filters: kernels such as unsharp masking or Sobel edge detection sharpen ridge flow and pore boundaries.

IV. Facial Reconstruction

Facial reconstruction rebuilds a likely appearance from skeletal or partial remains to generate investigative leads when conventional identification fails.

A. Purpose and Principle

• Tissue-depth premise: average soft-tissue thicknesses at defined cranial landmarks let muscle and skin be layered predictably over bone.

B. Facial Reconstruction

1. Manual (sculptural) method: tissue-depth pegs are fixed at landmarks (glabella, nasion, gnathion) on a skull cast, and clay is built up between them.
2. Computerised (3-D) method: the skull is CT-scanned, and software deforms a template face to the tissue-depth constraints, allowing rapid variant generation.

• Landmark anthropometry: roughly 21–34 standard points carry population- and sex-specific depth tables (e.g., thicker at cheeks in higher BMI groups).
• Feature estimation rules: nose width ≈ 1.67 × nasal aperture width; eyeball centred in the orbit; mouth width set by inter-canine distance.
• Superimposition link: the reconstructed or photographed face is overlaid on the skull to test consistency, tying reconstruction to impression-style comparison.

V. Analysis of Combination of Impression Evidences

Combining multiple impression types raises the collective evidential value beyond any single mark through corroboration and sequencing.

A. Rationale

• Cumulative probability: independent impressions (a shoeprint plus a fingerprint plus a lip mark) multiply the improbability of coincidental association with one individual.

B. Analysis of Combination of Impression Evidences

• Spatial correlation: mapping where each impression sits reconstructs movement and contact sequence at the scene.
• Sequencing (overlap analysis): which mark lies atop another establishes chronological order (a shoeprint over blood post-dates the bleeding).
• Cross-transfer confirmation: matching a tool mark to the tool and the tool's grip to a fingerprint chains suspect to weapon to scene.
• Weight aggregation: each discipline's likelihood ratio is combined so concordant results reinforce, while any discordant result flags re-examination.
  • Likelihood ratio (LR): probability of the evidence given a common source divided by probability given different sources; LR > 1 supports common source.

VI. Various Equipment Involved in Analysis

The reliability of impression conclusions depends directly on standardised capture and comparison hardware.

A. Overview

• Capture-to-comparison pipeline: equipment spans lighting, imaging, magnification and 3-D metrology.

B. Various Equipment Involved in Analysis

• DSLR/macro imaging systems: high-resolution sensors with macro lenses record fine minutiae at ≥ 1000 ppi for AFIS-quality prints.
• Comparison microscope: two stages fused through a bridge display two marks side-by-side for tool mark and striation matching.
• Video spectral comparator (VSC): switchable UV/visible/IR illumination with filters examines documents and treated latents.
• Alternate light source unit: multi-wavelength lamp for locating and photographing fluorescing residues.
• 3-D surface scanners: structured-light or laser profilometers digitise footwear and tool-mark relief for non-contact comparison.
• Electrostatic dust-print lifter (ESDL): applies high voltage to a metallised film, lifting dry-origin dust prints from floors.
• Gelatine and adhesive lifters: transfer developed marks from irregular surfaces for photographing under controlled light.

VII. TraX Software

TraX is a dedicated forensic image-processing platform built to enhance and manage impression images within a defensible workflow.

A. Purpose and Principle

• Audit-trail design: every filter, crop and adjustment is logged as a reversible step so the enhancement history is fully reconstructable in court.

B. TraX Software

• Non-destructive editing: the original image is preserved as a locked master; operations act on derived layers.
• Frequency-domain tools: integrated FFT pattern removal strips woven or printed backgrounds behind prints.
• Calibrated measurement: on-screen scaling from an embedded ruler yields metric footwear and tyre dimensions.
• Batch processing: consistent settings are applied across a series of exhibit images to keep comparisons uniform.
• Case management: images, metadata and step-logs are bundled per exhibit, supporting disclosure and peer verification.

VIII. SUT-Lip-DB: Database for Lip Prints

SUT-Lip-DB is a research reference database of lip-print (cheiloscopy) images assembled to develop and benchmark automated lip-print recognition.

A. Definition and Basis

• Cheiloscopic premise: the pattern of grooves on the labial mucosa (Suzuki–Tsuchihashi types I–V) is individual and, like fingerprints, persistent — the theoretical basis for treating lip prints as biometric impressions.

B. SUT-Lip-DB: Database for Lip Prints

• Structured samples: the collection stores multiple lip-print images per subject, enabling within-person and between-person comparison.
• Classification support: entries are labelled by groove type so algorithms learn the vertical, branched, reticular, and intersecting patterns.
  • Suzuki–Tsuchihashi grooves: Type I (clear vertical), I′ (partial vertical), II (branched), III (intersected), IV (reticular), V (undetermined).
• Feature extraction benchmarking: the dataset provides a fixed test bed for methods such as texture descriptors and minutiae-style groove matching.
• Automation goal: supports moving cheiloscopy from subjective visual comparison toward reproducible, quantifiable matching consistent with the ACE-V standard.
• Investigative value: lip prints left on cups, cigarettes or fabric become searchable evidence once digitised against such a reference set.