- Nationwide Digital Forensic & Cyber Investigation Services
Reverse projection is deceptively simple in concept and demanding in execution. Put the camera back where it was, put a measured reference where the subject stood, and compare. This page walks through the actual procedure, the assumptions it removes, and the ones it cannot.
| Question | One line answer |
|---|---|
| What makes reverse projection reliable? | The questioned image and the reference image share the same camera, lens, and viewpoint. |
| Is the original camera required? | Strongly preferred. An identical model with matched settings is the fallback. |
| What is the reference? | A measured scale, marked pole, or documented person placed at the subject position. |
| Does it need software? | The comparison is geometric. Software assists with overlay and documentation, it does not supply scale. |
| What is the output? | A height interval with the assumptions and error sources stated. |
| Can a jury see it? | Yes. The overlay is directly demonstrative, which is unusual among forensic measurements. |
| Term | What it means |
|---|---|
| Extrinsics | The camera position and orientation in the scene, which reverse projection reproduces physically rather than solving numerically. |
| Calibration reference | An object of documented dimensions placed at the subject location to establish scale in the image. |
| Overlay comparison | Superimposing the questioned frame and the reference frame so the subject and the scale can be read against each other. |
| Principal point | The image point where the optical axis meets the sensor, around which radial lens distortion is symmetric. |
| Camera matching | Solving a virtual camera position and lens in a 3D model to reproduce the original view without physical access. |
| Subject position | The precise scene location where the person stood, established by matching floor features between frames. |
Reverse projection is a scene method. Most of the work is done before anything is measured, and the quality of the setup determines the credibility of the result.
Verify the recording. Authenticate the export, document its properties, and confirm the frames to be used are unaltered.
Confirm camera continuity. Establish that the camera is the same unit, in the same mount, with the same lens and configuration as at the time of recording.
Fix the subject position. Use floor seams, tile lines, thresholds, and fixed fixtures to locate precisely where the subject stood.
Place the calibrated reference at that exact position, plumb and perpendicular to the floor plane, with its dimensions independently measured and photographed.
Record comparison imagery through the original camera, using the same resolution, frame rate, and settings as the questioned recording.
Overlay and measure. Compare the head top and floor contact landmarks of the questioned subject against the reference in the same image geometry.
Report an interval. Combine landmark uncertainty, posture variation, footwear, and positional error into a stated range with the reasoning shown.
The scene visit should produce a record that an opposing examiner can follow without ambiguity: photographs of the reference in place, measurements of the reference itself, notes on the camera settings, and imagery showing the correspondence between floor features in the questioned frame and the reference frame. If a second examiner cannot repeat the setup from the report, the analysis is not reproducible, and reproducibility is what a reliability challenge tests.
Once a camera is swapped or a floor is resurfaced, the strongest version of this method is gone. Early scene work protects the option.
Reverse projection removes the need to model the optics, but it does not remove the human variables. Being explicit about that division is the difference between a measurement and an overstatement.
Both images pass through the same lens, so distortion affects the questioned subject and the reference identically at the same scene position.
The physical camera supplies the true projection, so focal length does not have to be estimated.
The camera is physically in its original position rather than solved for from image cues.
Sole and heel thickness are recorded as part of the person and must be measured separately or reported as unresolved.
Stance, head tilt, slouching, and stride phase all move the upper landmark and are not correctable after the fact.
Compression, motion blur, and pixel density limit how exactly the head top and floor contact can be placed.
Peer reviewed comparison work has evaluated reverse projection against photogrammetric software approaches, and later work extended the method to scenarios with a moving camera and a moving subject. Both bodies of work reinforce the same point: the method is sound, and the reported uncertainty has to reflect the conditions of the specific case rather than the best case described in a textbook.
The method has hard prerequisites. When they cannot be met, the honest path is to change method and widen the interval, not to pretend the prerequisites were met.
| Obstacle | Effect on reverse projection | Alternative approach |
|---|---|---|
| Camera replaced with a different model | Optical path cannot be reproduced exactly | Characterize the new camera, or solve intrinsics and use camera matching |
| Scene renovated | Subject position and floor plane no longer exist | Use a prior laser scan, survey, or archived imagery if any exists |
| Camera moved or re-aimed | Viewpoint differs from the questioned frame | Attempt viewpoint restoration using fixed scene features, and report the residual error |
| No scene access at all | No physical reference can be placed | Single view metrology using in frame geometry, with a wider stated interval |
| Only images survive, no scene and no camera | Method unavailable | Statistical body height modeling with explicit uncertainty, or decline to give an opinion |
A report that switches methods should say plainly why, what was lost, and how the interval changed. That transparency is usually more persuasive to a court than a confident number of unclear provenance.
It removes optical modeling error. Footwear, posture, and landmark precision remain, so the output is still an interval.
A different sensor or lens produces a different projection. Substitution requires characterization and disclosure.
The overlay is the measurement, presented visually. Its persuasive quality comes from the underlying geometry, not from graphics.
Software approaches are legitimate and have been compared against reverse projection in the literature, but they reintroduce assumptions about the camera that physical reproduction avoids.
| Approach | Scale source | Key assumption | Typical use |
|---|---|---|---|
| Reverse projection | Physical reference at the subject position | Camera and scene reproduced faithfully | Scene and camera intact |
| Camera matching to 3D scan | Scanned scene geometry | Scan is dimensionally accurate and intrinsics solved correctly | Scene documented before change |
| Single view metrology | In frame geometry and a known feature | Parallel lines identified correctly, distortion corrected | No scene access |
| Multi image photogrammetry | Reconstructed geometry across images | Sufficient overlap and stable features | Multiple viewpoints available |
| Statistical modeling | Population data and image measurements | Subject falls within modeled population | Images only, uncertainty aware estimate |
Our examiners run reverse projection as a documented scene procedure, not as an afterthought to video review, and we build the record so the setup itself can withstand scrutiny.
Camera continuity verification, subject position location, calibrated reference placement, and full photographic documentation.
Frame by frame comparison of questioned and reference imagery with measured landmarks and documented selection reasoning.
Scene scanning where alteration is expected, preserving geometry for later camera matching.
A stated interval built from the identified error sources in the specific case rather than a generic margin.
Courtroom overlays and comparison boards that show the basis of the opinion clearly.
Examination of another examiner setup for camera substitution, position drift, or unmeasured references.
Consultations are confidential. We work with defense counsel, prosecutors, civil litigators, and investigative agencies nationwide, and we will tell you candidly when the footage cannot support a reliable measurement.
Elite Digital Forensics is an independent digital forensics firm serving defense attorneys, prosecutors, civil litigators, and investigative agencies nationwide. Our examiners include former state and federal law enforcement forensic examiners who have testified as court qualified expert witnesses. We are retained by either side of a matter, and our findings are reported the same way regardless of who retains us.
Every engagement follows documented chain of custody, reproducible measurement methodology, stated uncertainty, and reporting written for attorney review, negotiation, or courtroom use. Work performed at the direction of counsel is generally treated as attorney work product prepared in anticipation of litigation. Call (833) 292-3733 or request a confidential consultation.
It is a measurement method in which the original recording geometry is physically reproduced. The examiner returns the camera to its original position with matching optics, places a reference of known height where the subject stood, and compares the two images. Because both share the same optical path, most sources of modeling error cancel.
Software approaches must estimate camera parameters that reverse projection supplies physically. Both are legitimate and have been compared in peer reviewed testing, but reproducing the original geometry removes assumptions rather than modeling them, which is easier to defend on cross examination.
The examiner can characterize the replacement camera and correct for differences, or shift to camera matching against a 3D scan of the scene. Both are workable, and both widen the reported interval, which the report must state.
By matching fixed floor features between the questioned frame and the scene: tile grout lines, floor seams, thresholds, drains, and permanent fixtures. Positional error is one of the error sources folded into the final interval.
Yes, with more care. Published work has extended the method to cases where both the subject and the camera move. The analysis then requires multiple frames and a wider interval because stride phase changes apparent height.
No, though a documented stature for the person being compared is useful. The analysis measures the subject in the footage. Comparison to a known individual is a separate step, and it should account for footwear and time of day.
This content is for educational and informational purposes only and does not constitute legal advice. Elite Digital Forensics provides independent digital forensic services and expert witness testimony; we do not provide legal representation. Every case is fact specific; outcomes depend on the evidence, jurisdiction, and counsel. Retain qualified legal counsel for advice about your matter.
Elite Digital Forensics Assistant
By submitting this form, you consent to be contacted by email, text, or phone. Your information is kept secure and confidential. Reply Stop to opt out at anytime.
IMPORTANT: Please remember to check your spam or junk folder