# Spotting Cheetahs: Identifying Individuals by Their Footprints

> Can you tell one cheetah from another by their footprints alone? Using photographs of 781 footprints from 38 captive cheetahs, researchers built a footprint identification technique that told individuals apart with over 90% accuracy. They argue that skilled trackers will be the key to taking it into the wild.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/2016-jewell-alibhai-spotting-cheetahs-identifying-individuals/

## Citation and access

- **Authors:** Zoe C. Jewell, Sky K. Alibhai, Florian Weise, Stuart Munro, Marlice Van Vuuren and Rudie Van Vuuren
- **Year published:** 2016
- **Journal or source:** Journal of Visualized Experiments
- **Full citation:** Jewell, Z. C., Alibhai, S. K., Weise, F., Munro, S., Van Vuuren, M. and Van Vuuren, R. 2016. Spotting cheetahs: identifying individuals by their footprints. Journal of Visualized Experiments (111): e54034.
- **DOI:** 10.3791/54034
- **Access status:** Open access (CC BY-NC-ND 3.0)
- **PDF:** https://www.originalwisdom.com/wp-content/uploads/2026/11/2016_Jewell_and_Alibhai_etal_spotting-cheetahs-identifying-individuals-by-their-footprints.pdf

## Tags

- **Species:** Mammals > Cheetah
- **Region:** Africa > Namibia
- **Tracking:** Individual identification; Track measurements; Track photographs; Tracking; Tracks
- **Methods and technology:** Footprint Identification Technology (FIT)
- **Data type:** New field data; Protocol or guidelines
- **Observer qualification:** Not reported
- **Discipline:** Wildlife tracking
- **Year published:** 2016

## Detailed keywords

- **Species:** Studied: cheetah (Acinonyx jubatus), 38 registered captive individuals (16 females, 22 males, aged 1 to 11 years). Earlier uses of the technique mentioned: black rhinoceros (Diceros bicornis), white rhinoceros (Ceratotherium simum), Amur tiger (Panthera tigris altaica), mountain lion, lowland tapir.
- **Location and habitat:** Captive cheetahs walking over prepared sand trails, about 2 to 3 m wide and 3 to 15 m long, laid along a perimeter fence or habitual path. The paper does not say where the footprints were collected. The work is set in the context of cheetah conservation in Namibia, with the N/a’an ku sê Foundation and Research Programme (Namibia); the acknowledgements also thank Chester Zoo (United Kingdom), Cheetah Conservation Botswana and Foundation SPOTS (Netherlands).
- **Topics:** footprint identification technique (FIT); individual identification from footprints; non-invasive monitoring; standardized footprint photography protocol; landmark points and morphometrics; training dataset of known individuals; holdback validation; misclassification between littermates; tracking skills needed for free-ranging animals; coalitions of males moving together; citizen science and smartphones; farmer and cheetah conflict; cost of collars and camera traps.
- **Statistics or models:** robust cross-validated pairwise discriminant analysis with a customized model and a Reference Centroid Value (RCV); Ward’s cluster analysis (dendrograms); holdback trials (test sets of 4, 8, 12 and so on, 10 random iterations each); 136 variables (distances, angles, areas) from 25 landmark and 15 derived points per footprint; run as an add-in to JMP data visualization software (JMP Scripting Language).
- **Data type:** New field data from captive cheetahs (781 left hind footprints); no dates of data collection given. A step-by-step protocol with a video article. Inconsistency: the text says 110 trails, but Table 1 totals 109 trails (55 from females, 54 from males); the 5,886 pairwise comparisons reported match 109 trails, not 110. Step 1.3.5 refers to steps “1.1.6 to 1.3.5”, but section 1.1 has only four steps.

## Synopsis

### 1. What they did

Cheetahs are Africa’s most endangered large cat, and Namibia holds perhaps more than a third of the world’s cheetahs. Estimates of their numbers vary widely because different monitoring methods are used, and collars and camera traps are costly or limited. The authors set out, step by step and with a video, how to identify individual cheetahs from photographs of their footprints, using the footprint identification technique (FIT) first developed for black rhinoceros.

- Collecting footprints. Captive cheetahs were lured with a food reward across a raked, wetted sand path. Only left hind prints were used. Each print was circled, then photographed from directly overhead with a metric scale and a label slip, about 20 good prints per animal, early or late in the day for good contrast.

- Measuring the prints. In the software, each image was rotated to a standard position and scaled, and 25 landmark points were placed by hand on the toes and heel. The software added 15 derived points and calculated 136 measurements (distances, angles and areas) per print.

- Building the model. Trails were compared in pairs to decide whether two trails came from the same cheetah or from two different cheetahs. The results were drawn as a cluster tree (dendrogram), and the number of measurements, the confidence ellipse size and the cut-off were adjusted until the model best matched the known animals.

- Testing it. In holdback trials, groups of cheetahs (4, 8, 12 and so on) were hidden from the model and their number predicted, 10 times for each group size.

### 2. What they found

- The training set held 781 left hind footprints from 38 cheetahs: 386 prints from 16 females and 395 from 22 males, aged 1 to 11 years.

- In the holdback trials the predicted number of cheetahs stayed close to the true number, even when the test set (28 cheetahs) was much larger than the training set (10 cheetahs).

- Accuracy was consistently above 90%, both for the number of individuals and for classifying trails.

- With all trails included there were 5,886 pairwise comparisons and 46 misclassifications, an overall accuracy of 99%. Trails from the same cheetah were matched correctly 93% of the time (117 of 126); trails from different cheetahs were kept apart 99% of the time (5,723 of 5,760).

- Many of the misclassifications were between littermates.

- In a sample of seven cheetahs, the best settings (18 measurements) predicted seven animals correctly; other settings (24 or 10 measurements) gave wrong numbers.

### 3. What they concluded

- FIT is a cost-effective, non-invasive and community-friendly way to monitor cheetahs, needing only a basic digital camera and a scale. It can be used to count individuals, follow known animals, or supply “marks” for mark-recapture density estimates.

- What sets it apart from earlier footprint methods is the standardized collection protocol, simple software, standard orientation of images, and the new statistical model.

- Critical steps: a well-prepared sand trail with the animal walking at a relaxed pace, the camera directly over the print, and a photographer or assistant (perhaps an expert tracker) who can recognise a cheetah footprint and follow the trail.

- Placing landmark points by hand is central to accuracy; for now one operator per site should do it, and automation is being developed.

- Next steps: field trials with free-ranging cheetahs, algorithms for age class, substrate controls, computer vision, and a smartphone app so that visitors to protected areas could collect footprint data.

### 4. How tracking was or was not used

The study used footprints of captive cheetahs walking over prepared sand, so no trailing of wild animals took place. The operators had to tell left hind prints from front prints (the leading third toe and the slope of toes 3, 4 and 5 to the left; front feet are broader) and place landmark points on the toes and heel. The authors stress that tracking skills are essential for using the technique on free-ranging cheetahs: a tracker must identify cheetah footprints, tell them apart from other large carnivores, and follow a trail forwards or backwards so that trails of animals travelling together, such as male coalitions, are not mixed up. They note this was solved for white rhinoceros groups of up to 13 animals by tracking trails forwards or backwards.

### 5. Tracker or observer certification

Not reported. The paper does not say who collected, photographed or marked the footprints. It mentions the rapid growth of tracker training and certification programs, and the planned Academy of Ancient Skills at the N/a’an ku sê Foundation for passing on the skills of expert indigenous trackers.

### 6. Other technologies, skills or innovations

- The FIT add-in for JMP data visualization software, with a graphic user interface for feature extraction, pairwise analysis and holdback trials.

- A depth correction for prints in substrate deeper than 1 cm.

- A Reference Centroid Value (a duplicated copy of the database) that stabilizes the pairwise comparisons.

- A smartphone app being developed for WildTrack.

- The publication fees were paid by the SAS Institute, which makes JMP.

### 7. Implications for conservation (from the paper)

- Reliable data on cheetah numbers, identity and distribution are key to dealing with conflict between farmers and cheetahs in Namibia, where more than 1,200 cheetahs were removed from 1991 to 2006.

- Because FIT is non-invasive, it avoids the risks and costs of capture, drugging and collaring.

- Built on traditional tracking skills and low cost, it can involve previously marginalized local communities in conservation monitoring.

- Artificial sand trails could be used with camera traps, for example at cheetah marking trees, where natural ground is unsuitable.

### 8. Implications for tracking (Original Wisdom's observations)

_These are Original Wisdom's observations, not the authors'._

- The accuracy figures come from captive cheetahs walking on prepared sand with known identities. On natural ground, the quality of the prints and the tracker’s ability to keep each animal’s trail separate will matter much more, as the authors themselves point out.

- Misclassifications clustered among littermates, which suggests related cheetahs may have similar feet; trackers working with family groups or coalitions may need extra care.

## Cite this summary

Lawrence, K. 2026. Summary of: Spotting Cheetahs: Identifying Individuals by Their Footprints (Jewell and Alibhai et al., 2016). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/2016-jewell-alibhai-spotting-cheetahs-identifying-individuals/