Fig. 1. Comparison of imaging depth and resolution between optical, ultrasound and photoacoustic imaging^[15-19] (CFM: confocal microscopy; TPM: two-photon microscopy; OCT: optical coherence tomography; OR-PAM: optical-resolution photoacoustic microscopy; AR-PAM: acoustic-resolution photoacoustic microscopy; PACT: photoacoustic tomography; USI: ultrasound imaging)

Fig. 2. Skin structure, photoacoustic imaging principles and combined diagrams of vascular imaging. (a) Human skin structure^[20] (Ep: epidermis, 0.05‒1.5 mm; De: dermis; 1‒4 mm, up to 6 mm thick; Hy: hypodermis, thicknesses vary from several millimeters to several centimeters); (b) photoacoustic imaging principle diagram^[21]; (c) different depths of human palm skin were imaged using 532 nm and 1064 nm wavelength laser as excitation sources for photoacoustic microimaging^[24], where images Ⅰ‒V corresponding to depths of 0‒0.1 mm, 0.1‒0.25 mm, 0.25‒0.72 mm, 0.72‒1.8 mm and 1.8‒3 mm, respectively

Fig. 3. Schematic diagrams of photoacoustic microscopy imaging system, laser illumination mode, and focusing objective lens. (a) OR-PAM system with optical focusing capability and AR-PAM system with acoustic focusing capability^[30], (i) for the OR-PAM component, (ii) for the AR-PAM component; (b) schematic diagram of laser beam focusing for a bifocal alternately illuminated photoacoustic microscopy system^[22]; (c) overall architecture of fast controlled confocal focus PAM system^[42]; (d) a noninvasive photoacoustic microbiopsy system based on an adjustable confocal photoacoustic objective lens^[43], schematic diagram of photoacoustic signals generated by multilayered skin tissues on the left, structure of the system objective lens on the right

Fig. 4. Schematic diagrams of the ultrasound transducers in several photoacoustic microscopy imaging systems. (a) Curved ultrasonic transducer^[47]; (b) ring ultrasound transducer^[47]; (c) hemispherical ultrasonic transducer^[47]; (d) Fabry-Perot based ultrasonic transducer^[48]; (e) linear piezoelectric micromachined ultrasonic transducer^[57]; (f) dual piezo chip transducer^[58]; (g) hollow bowl ultrasonic transducer^[52]

Fig. 5. Structural diagram of hand vascular photoacoustic skin tomography system^[68]

Fig. 6. Schematic diagrams of laser irradiation modality and ultrasound transducers of several other photoacoustic skin imaging systems. (a) Manually controlled multi-angle lighting technology^[80], the left side is the probe structure and the right side is the diagram of adjusting irradiation angle of target area ; (b) ring ultrasound transducer^[28]; (c) linear array ultrasound transducer^[28]; (d) hemispherical ultrasonic transducer^[28]; (e) hyperbolic array transducer^[83]; (f) hemispherical rounded top hat for placement of sparse single crystal probe ultrasound transducer^[85]; (g) optical cage irradiation method^[79]

Fig. 7. Structures of several deep learning methods for improving imaging resolution. (a) NETT code/decode training strategy framework diagram^[112]; (b) FD Unet dense block detail showing diagram^[117]; (c) schematic diagram of Deep-E training and workflow^[118-119]

Fig. 8. Structure of the four-dimensional spectral spatial imaging computational method^[130]. (a) Data acquisition sets the relevant parameters and displays the distribution of the model light field and detector acoustic field under a collimated Gaussian beam in the model; (b) relevant performance optimization parameters; (c) experimental system; (d) training spread-spectrum network model; (e) training depth-enhanced network model; (f) spread-spectrum network model output results; (g) dept-enhanced network model output results

Fig. 9. Structures of several noise removal methods. (a) Flowchart of photoacoustic signal recovery algorithm^[140]; (b) As-Net network structure diagram^[146]; (c) unsupervised domain change network framework diagram^[147]

Fig. 10. Photoacoustic skin imaging application in skin cancer disease. (a) Melanomas, where image Ⅰ is 3D image of melanomas^[159], image Ⅱ is statistical graph quantifying the number of blood vessels and maximum thickness of melanomas over time of tumor development^[160], image Ⅲ is photoacoustic imaging images of melanomas in mice on the seventh day^[160], and image Ⅳ is photoacoustic imaging image of melanomas in mice on the thirtieth day^[160]; (b) squamous cell carcinoma of the skin, where image Ⅰ is photoacoustic spectrum of squamous cell carcinoma of the skin and its surrounding healthy skin (the green line represents healthy skin and the red line represents squamous cell carcinoma, and the difference between healthy and diseased tissues can be clearly seen in the wavelength range of 765‒960 nm)^[165], image Ⅱ is the comparison of each substance's content in squamous cell carcinoma diseased tissues with that in the healthy tissues^[164], and image Ⅲ is imaging of the diseased area at the same location of the ear of the mice on the third and the ninth day, respectively (vascular proliferation can be more clearly observed)^[69]; (c) basal cell carcinoma, where images Ⅰ and Ⅱ show the boundary between diseased and healthy tissues can be clearly distinguished by photoacoustic spectroscopy^[170], and image Ⅲ is vascular morphology comparison between healthy tissue and basal cell carcinoma lesions in the mouse ear^[156]

Fig. 11. Photoacoustic skin imaging application in other dermatologic conditions. (a) Comparative imaging of human healthy skin and psoriatic lesion tissue^[181], where image Ⅰ is healthy skin imaging, image Ⅱ is psoriatic lesion tissue imaging; (b) comparative imaging of human healthy skin and nevus erythematosus lesion tissues^[43],where image Ⅰ is healthy skin imaged with normal number of blood vessels, image Ⅱ is nevus erythematosus lesion skin imaged with a significant increase in the number of microvessels in the subcutaneous area, and image Ⅲ is microvessel density statistic for both healthy skin and lesion tissues; (c) comparative imaging of human healthy skin and café au lait lesion tissues^[46,53], where image Ⅰ shows the basal layer thickness of healthy skin, image Ⅱ shows basal layer thickness of lesion skin, image Ⅲ shows statistical comparison of melanin content and basal layer thickness between healthy and café au lait skin (blue bars represent healthy skin and red bars represent café au lait skin) and image Ⅳ shows the distribution of photoacoustic amplitude along the direction of epidermal depth in healthy skin and café au lait skin